1 NON-INNOCENCE IN 2Fe2S CLUSTERS FOR THE PURPOSE OF MOLECULAR HYDROGEN PRODUCTION

1  NON-INNOCENCE IN 2Fe2S CLUSTERS FOR THE PURPOSE OF MOLECULAR HYDROGEN PRODUCTION
1
NON-INNOCENCE IN 2Fe2S CLUSTERS FOR THE PURPOSE OF MOLECULAR
HYDROGEN PRODUCTION
by
Gabriel B. Hall
____________________________
A Dissertation Submitted to the Faculty of the
DEPARTMENT OF CHEMISTRY AND BIOCHEMISTRY
In Partial Fulfillment of the Requirements
For the Degree of
DOCTOR OF PHILOSOPHY
WITH A MAJOR IN CHEMISTRY
In the Graduate College
THE UNIVERSITY OF ARIZONA
2014
2
THE UNIVERSITY OF ARIZONA
GRADUATE COLLEGE
As members of the Dissertation Committee, we certify that we have read the dissertation
prepared by Gabriel B. Hall, titled Non-Innocence in 2Fe2S Clusters for the Purpose of
Molecular Hydrogen Production and recommend that it be accepted as fulfilling the
dissertation requirement for the Degree of Doctor of Philosophy.
_________________________________________________ Date: 4th February 2014
Professor Dennis L. Lichtenberger
_________________________________________________ Date: 4th February 2014
Professor Richard S. Glass
_________________________________________________Date: 4th February 2014
Professor Zhiping Zheng
_________________________________________________Date: 4th February 2014
Professor John H. Enemark
_________________________________________________Date: 4th February 2014
Professor Douglas A. Loy
Final approval and acceptance of this dissertation is contingent upon the candidate’s
submission of the final copies of the dissertation to the Graduate College.
I hereby certify that I have read this dissertation prepared under my direction and
recommend that it be accepted as fulfilling the dissertation requirement.
____________________________________________________Date: 4th February 2014
Dissertation Director: Professor Dennis L. Lichtenberger
3
STATEMENT BY AUTHOR
This dissertation has been submitted in partial fulfillment of requirements for an
advanced degree at the University of Arizona and is deposited in the University Library
to be made available to borrowers under rules of the Library.
Brief quotations from this dissertation are allowable without special permission,
provided that accurate acknowledgment of source is made. Requests for permission for
extended quotation from or reproduction of this manuscript in whole or in part may be
granted by the head of the major department or the Dean of the Graduate College when in
his or her judgment the proposed use of the material is in the interests of scholarship. In
all other instances, however, permission must be obtained from the author.
SIGNED: Gabriel B. Hall
4
Contents
List of Figures ..................................................................................................................... 7
List of Tables .................................................................................................................... 11
List of Schemes ................................................................................................................. 12
List of Abbreviations ........................................................................................................ 12
Abstract ............................................................................................................................. 13
Chapter 1 Introduction ...................................................................................................... 17
Natural Hydrogen Production ....................................................................................... 17
Hydrogenase Inspired Complexes ................................................................................ 18
General Background ................................................................................................. 18
Implications of Non-innocence ................................................................................. 24
Utilizing Non-Innocent Chromophores to Harvest Light. ........................................ 26
Cyclic Voltammetry .................................................................................................. 30
Density Functional Theory Calculations ...................................................................... 34
Chapter 2 Experimental Methodologies ........................................................................... 35
Electrochemistry ........................................................................................................... 35
Notes to Future Team DLL Students on Performing Cyclic Voltammetry .............. 36
Preparing Solutions ................................................................................................... 36
Preparing the Cell and Electrodes ............................................................................. 37
Estimating Solution Resistance................................................................................. 41
Performing a 2-segment Scan ................................................................................... 42
3-segments Scans ...................................................................................................... 47
Fc+/Fc Correction and Backgrounds ......................................................................... 48
Data Work-Up ........................................................................................................... 49
EPR Spectroscopy......................................................................................................... 49
Computational Methods ................................................................................................ 49
Infra-Red Spectorelectrochemistry ............................................................................... 50
Notes to Future Team DLL Students on Performing IR-SEC .................................. 51
Chapter 3 Redox Active Quinones Coupled to 2Fe2S Cores Pertaining to Molecular
Hydrogen Production ........................................................................................................ 52
Introduction: .................................................................................................................. 52
Results and Discussion ................................................................................................. 55
Structure. ................................................................................................................... 55
5
Electrochemistry. ...................................................................................................... 58
EPR Spectroscopy......................................................................................................... 65
Computations ................................................................................................................ 68
Conclusions ................................................................................................................... 71
Chapter 4 Comparison of the Electronic Structure of 1,2-(µ-benzenedithiolato)-2’phenylazopyridinediiron-tetracarbonyl and 1,3-(µ-propanedithiolato)-2’phenylazopyridinediiron-tetraacarbonyl to 1,2-(µ-benzenedithiolato)-hexacarbonyl and
1,3-(µ-propanedithiolato)-hexacarbonyl ........................................................................... 74
Introduction ................................................................................................................... 74
Comparison of structural parameters: ........................................................................... 77
IR spectroscopy: CO stretching frequencies ................................................................. 77
UV-Vis spectroscopy .................................................................................................... 80
DFT computations and correlation diagrams ................................................................ 82
Conclusions ................................................................................................................... 90
Chapter 5 Persistence of Disulfide Bonds in Reduced Bipyridines and the Observance of
Intramolecular Charge Transfer. ....................................................................................... 91
Introduction ................................................................................................................... 91
Results and Discussion ................................................................................................. 96
Behavior of 17........................................................................................................... 96
Behavior of 19a2+ .................................................................................................... 106
Conclusions ................................................................................................................. 114
Experimental ............................................................................................................... 115
Chapter 6 Asymmetrical Diiron Bisthiolate Complexes, [(–SR)(–SR)Fe2(CO)6], for
Electrocatalytic Proton Reduction from Weak Acid ...................................................... 117
Introduction ................................................................................................................. 117
Results and Discussions .............................................................................................. 119
Synthesis and Structural Analysis ........................................................................... 119
Electrochemistry ..................................................................................................... 128
Gas-phase Photoelectron and UV-Vis Spectroscopy .............................................. 133
Electrocatalytic Proton Reduction of Weak Acid ................................................... 136
Conclusions ................................................................................................................. 139
Experimental Section .................................................................................................. 140
General Procedures. ................................................................................................ 140
X–ray Structure Determination ............................................................................... 140
Electrochemistry ..................................................................................................... 141
IR spectroelectrochemistry ..................................................................................... 141
6
Gas–phase UV Photoelectron Spectroscopy ........................................................... 142
Density Functional Theory Calculations ................................................................ 142
Chapter 7 Conclusions and Future Directions ................................................................ 143
References ....................................................................................................................... 149
7
List of Figures
Figure 1.1 The [FeFe]-hydrogenase active site, shown with synthetic variants inspired by
[FeFe]-hydrogenase. ......................................................................................................... 21
Figure 1.2. (a) ground state of a neutral catalyst with attached chromophore. (b) excited
state of a neutral catalyst that is in the process of being reduced at a lower energy due to
electronic excitation. (c) reduced catalyst ......................................................................... 27
Figure 1.3 Two examples of [Fe-Fe]-hydrogenase inspired molecules designed for light
harvesting. ......................................................................................................................... 28
Figure 1.4 Cyclic voltammogram of a dichloromethane solution of 1.08 mM ferrocene,
and 0.2 M tetrabutylammonium hexafluorophosphate on a glassy carbon electrode
showing both empirical data (green line) and simulated voltammogram (black circles). 33
Figure 2.1 Electrochemical cell for cyclic voltammetry employed by our laboratory
showing A) Pt counter electrode with air vent. B) Glassy carbon working electrode. C)
Inert air bubbler. D) AgNO3/Ag counter electrode with salt bridge. Below shows
electrode connectivity. ...................................................................................................... 38
Figure 2.2 A) Experiments tab for Gamry Framework. B) Graph of a generic cyclic
voltammogram displaying Potential Vs. Time ................................................................. 43
Figure 2.3 Gamry Framework display window and settings used to perform cyclic
voltammetry on 1 mM Fc in acetonitrile. ......................................................................... 44
Figure 3.1 Quinone complexes examined in this study. ................................................... 54
Figure 3.2 Single crystal structure of 8b, and 8c with displacement ellipsoids at the 50%
probability level. The structure has two–fold rotational symmetry; the asymmetric unit is
the half of the molecule which is numbered. .................................................................... 56
Figure 3.3 Organic quinones in this study. ....................................................................... 57
Figure 3.4 Voltammograms of ca. 0.5 mM quinone complexes 7b, 8a and 10a as well as
the corresponding quinones: 11b, 12 and 13. The currents have been divided by the
concentrations of the compounds to normalize for slight variations in concentration.
Dichloromethane with 0.10 M n–Bu4NPF6 glassy carbon working electrode, 0.10 V/s. . 60
Figure 3.5 Voltammograms of 0.48 mM 8a with additions of 0, 1 and 5 mM acetic acid.
Mercury–film working electrode. Other conditions as in Figure 3.4. .............................. 64
8
Figure 3.6 Simulated (green) spectrum overlaid with the experimental (black) spectrum
for 11b. The deviation from agreement can be seen in blue. ............................................ 66
Figure 3.7 EPR Spectra for the semiquinone obtained on electrochemical reduction of 7be, 8a, 8b............................................................................................................................. 67
Figure 3.8 Calculated SOMO of 8b—. .............................................................................. 72
Figure 4.1 Comparison of the experimental IR spectrum (black) and calculated spectrum
(blue) in the metal-carbonyl stretching frequency region for 15 (above) and 16 (below).78
Figure 4.2 A.UV-Vis spectrum of 15 (blue) 4 (purple) and 14 (red) in dichloromethane.
B.UV-Vis spectrum of 16 (blue) 2a (purple) and 14 (red) in dichloromethane. .............. 81
Figure 4.3 LUMO +1 (top) LUMO (middle) and HOMO (below) calculated for 4 (left)
and 15 (right)..................................................................................................................... 84
Figure 4.4 Orbital energy correlation diagram. (a) Orbital energies of parent
hexacarbonyl compound (the “2Fe2S” is a block of seven orbitals that derive from the
two d6 Fe centers and the Fe-Fe bond). (b) The “2Fe2S” fragment orbital energies. (c)
Orbital energies of overall compound with the numbers leading to the LUMO showing
the percent contribution of contributing fragment orbitals (d) LUMO of 14. .................. 85
Figure 4.5 LUMO +1 (top) LUMO (middle) and HOMO (below) calculated for 3a and
16....................................................................................................................................... 88
Figure 5.1 Cyclic voltammograms of 0.97 mM 17 in 0.10 M Bu4NPF6/acetonitrile at 295
K. Glassy carbon working electrode. ................................................................................ 97
Figure 5.2 Experimental cyclic voltammogram of 17 under the same conditions as
present in Figure 5.1, and simulation (black circles) according to Scheme 5.2. E4 = -1.35
V α4= 0.75, ks4= 0.003 cm/s, E5 = −0.90 V; 5 = 0.66; Dall species = 2.3  10−5 cm2/s; rdisk
= 0.15 cm. Subscript numbers refer to the chemical equations in Scheme 5.2. The
simulation is based on diffusion to a disk electrode. ........................................................ 98
Figure 5.3 Experimental cyclic voltammogram of 17 under the same conditions as
present in Figure 5.1, and simulation (black circles) according to E6 = -1.435 V α6=
0.896, ks6= 0.052 cm/s, E8 = -0.891 V α8= 0.587, ks8= 0.003 cm/s, Keq7= 5.93, kf7=
11650, Dall species = 2.3  10−5 cm2/s; rdisk = 0.15 cm. Subscript numbers refer to the
chemical equations in Scheme 5.2. The simulation is based on diffusion to a disk
electrode. ........................................................................................................................... 99
9
Figure 5.4 Energy rotation profile, with respect to the S-C4-C4’-S’ torsion angle for 17
(top trace), 17- (middle trace) and 172- (bottom trace). ................................................... 104
Figure 5.5 LUMO of 17 (top), SOMO of 17- (center) and HOMO of 172- (bottom). .... 105
Figure 5.6 Cyclic voltammogram of 0.73 mM 19a2+ in 0.10 M Bu4NPF6/acetonitrile at
298 K. Glassy carbon working electrode. Scan rate: 1.00 V/s. Line:
Background−corrected experimental cyclic voltammogram. Symbols: simulation for four
successive reversible electron−transfer reactions. The reactions were treated as reversible
(ks−values set at 0.3 cm/s). Standard potentials: E1 = −0.49; E2 = −0.60; E3 = −1.79;
E4 = −2.05 V.Dall species = 1.9  10−5 cm2/s..................................................................... 108
Figure 5.7 Energy rotation profile, with respect to the S-C4-C4’-S’ torsion angle for
19a2+ (top trace), 19a+ (second from top), 19a (third from top) 19a- (second from
bottom), and 19a2- (bottom trace). .................................................................................. 110
Figure 5.8 LUMO of 19a2+ (top), SOMO of 19a+ (middle), and HOMO of 19a (bottom).
......................................................................................................................................... 112
Figure 5.9 Cyclic voltammogram of 0.29 mM 18+ in 0.10 M n-Bu4NPF6/acetonitrile at
298 K. Glassy carbon working electrode. Scan rate: 1.0 V/s. The vertical bars at top are
the first, second, and third reduction potentials from the DFT computations. Note the
potential inversion of the first two reduction potentials. ................................................ 113
Figure 6.1 Thermal ellipsoid plots of molecular structures of 23a, 24a, b, and an overlay
of the crystal structures of 23a and, 24a, all viewed along the Fe-Fe bond. Display of 24a
and 24b is half the asymmetric unit. Thermal ellipsoid plots at 50% probability,
Hydrogen atoms have been omitted for clarity. .............................................................. 121
Figure 6.2 LUMO through HOMO – 2 of 23a and 24a.................................................. 124
Figure 6.3 HOMO – 3 through HOMO – 6 of 23a and 24a. .......................................... 125
Figure 6.4 HOMO – 7 through HOMO – 10 of 23a and 24a. ........................................ 126
Figure 6.5 Anodic and cathodic scans of 23a (solid line) and 24a (dashed line) in CH3CN
under N2 atmosphere. Arrows indicate the direction of scans. ....................................... 127
Figure 6.6 IR spectra in CO region of 23a (solid line) and 24a (dashed line) in mineral
oil. ................................................................................................................................... 130
10
Figure 6.7 IR-SEC spectra obtained by reduction of 23a (top) at -1.4 V and 24a (bottom)
at -1.2 V Vs. a pseudo silver reference in an Ar saturated solution of CH3CN. The solid
line indicates the spectrum post-electrolysis and dashed line indicates the spectrum preelectrolysis ...................................................................................................................... 131
Figure 6.8 He I (solid black) and He II (dashed red) ultra-violet photoelectron spectra of
23a (top) and 24a (bottom). ............................................................................................ 134
Figure 6.9 UV-Vis absorbtion spectra of 0.2 mM 23a (solid line) and 24a (dashed line) in
hexane solution. .............................................................................................................. 135
Figure 6.10 Voltammograms of ca. 1 mM 23a (top) and 24a (bottom) in 0.10 M
Bu4NPF6/CH3CN at 0.10 V/s in the absence and presence of various concentrations of
HOAc. Red arrow indicates initial direction of potential sweep. Return waves have been
omitted for clarity ........................................................................................................... 138
11
List of Tables
Table 3.1 Standard potentials of quinone complexes in comparison to the corresponding
quinones. ........................................................................................................................... 62
Table 3.2 EPR Spectroscopic Parameters of Complexed and Uncomplexed
Semiquinones. ................................................................................................................... 70
Table 3.3 Calculated EPR Spectroscopic Hyperfine Splitting and Mulliken Spin
Densities............................................................................................................................ 73
Table 4.1 Comparison of key bond lengths for 15 and 16 between the X-ray crystal
structure and DFT calculations. ........................................................................................ 79
Table 4.2 Calculated solvated energies (eV) in CH2Cl2, CH3CN and CH3CH2OH, of
various singlet excited states to the LUMO of compound 15........................................... 83
Table 4.3 Calculated solvated energies (eV) in CH2Cl2, CH3CN and CH3CH2OH, of
various singlet excited states to the LUMO of compound 16........................................... 83
12
List of Schemes
Scheme 5.1 Relevant compounds to this chapter.............................................................. 92
Scheme 5.2 Proposed electrochemical mechanism for compound 17. ........................... 101
Scheme 5.3 "Conventional" mechanism of S-S bond cleavage. ..................................... 102
Scheme 6.1 (a) THF, 2-ThMgBr, -78 ºC, 30 min; (b) CH3I, -78 ºC, 30 min; ambient
temperature, 20 hr. .......................................................................................................... 120
List of Abbreviations
Abbreviation
ADF
C
CV
DFT
E
Fc+/Fc
GGA
HOMO
Ipa
Ipc
IR
LDA
LUMO
OPBE
Ph
PPh3
Th
TZP
VWN
ZORA
Definition
Amsterdam Density Functional
A chemical step in an electrochemical mechanism
Cyclic voltammogram
Density functional theory
A transfer of an electron in an electrochemical mechanism
Ferrocenium/Ferrocene redox couple
Generalized gradient approximation
Highest occupied molecular orbital
Current at the maximum anodic peak potential
Current at the maximum cathodic peak potential
Infrared
Local density approximation
Lowest unoccupied molecular orbital
Optimized Perdew-Becke exchange
phenyl
Triphenylphosphine
thienyl
Triple-zeta potential
Vosko-Wilk-Nusair
Zeroth-order relativistic approximation
13
Abstract
The production of molecular hydrogen as a fuel source to replace traditional
carbon-based sources is key to the world’s environment and the national security of the
United States. Production of molecular hydrogen by energy efficient means with
abundant (inexpensive) materials is a recently emerged field throughout science. One
approach is to look to nature for inspiration, and this is indeed what our research group
has done by mimicking the active site of [FeFe]-hydrogenases. Unfortunately all current
models operate at a potential considerably more negative than the thermodynamic
potential of hydrogen production. This work attempts to address this deficiency by
modulating the reduction/oxidation potentials of the 2Fe2S core by binding it to noninnocent ligands. Non-innocent ligands are ligands which have redox behavior coupled to
the reduction/oxidation events of the metal center to which they are bound. This includes
ligands which are chromophores, potentially allowing for the reduction of excited states
at a less negative potential than the ground state of the complex.
One group of catalysts examined in this work is a series of substituted (μ-S2-1,4quinone) Fe2 (CO)4L2 complexes where L= CO, or PR3. The electronic communication
between the disulfide ligands and the metal centers where catalysis occurs dictates the
reactivity of a catalytic complex, and thus knowledge of this interaction is vital to
construct more efficient catalysts for hydrogen production. These complexes are found to
go through two separate one-electron reductions analogous to their parent 1,4-quinone
compounds but with much less negative reduction potentials. The once-reduced quinone
complex is relatively stable and gives the ability to study the electronic communication of
the radical species via electron paramagnetic spectroscopy (EPR). Delocalization to the
14
iron centers is demonstrated visibly by the 31P hyperfine splitting in the EPR spectrum of
a phosphine-substituted derivative. Modeling these EPR spectra with DFT calculations
indicates about 20% spin electron delocalization from the semiquinone anion radical to
the iron centers, and changing the functionality of the quinone gives the ability to tune
spin density at the metal centers. In the presence of excess acid, the electrochemically
produced semiquinone reacts to form the hydroquinone derivatives which subsequently
form molecular hydrogen.
One possible way of lowering the overpotential of hydrogen producing catalysts
is to use a chromophore to capture the energy contained in light, and then transfer the
energy to the catalyst active site. One example of a chromophore, which also happens to
be capable of binding metal centers is 2-phenylazopyridine. In order to study the
interplay between this chromophore and 2Fe2S catalysts, of the study of two compounds,
1,2-(µ-benzenedithiolato)-2’-phenylazopyridinediiron-tetracarbonyl and 1,3-(µpropanedithiolato)-2’-phenylazopyridinediiron-tetraacarbonyl, was undertaken. The UVVis spectra of both complexes show an intense absorption with a molar extinction
coefficient in the range of a ligand-to-metal charge transfer; however the wavelength of
maximum absorption does not show a dependence on solvent polarity for either complex.
Time-dependent-DFT calculations predict the UV-Vis spectra well and show the
transition to indeed be a ligand-to-metal charge transfer. DFT calculations show that the
difference between the energy of the ground state and excited state has little variance
with solvent polarity. Additionally molecular orbital correlation diagrams are constructed
to illustrate the relative orbital energies and intramolecular interactions of the complex.
15
Of particular interest is that the LUMO of 2-phenylazopyridine is relatively low lying in
energy, and mixes to comprise approximately 65% of the LUMO of each complex.
Modeling the electronic structure of a non-innocent ligand when not ligated can
aid in understanding the electronics of a complex containing the ligand. With the aim of
eventually synthesizing a series of [FeFe]-hydrogenase mimics with non-innocent 4,4’bipyridine 3,3’-dithiolato ligands; the behavior upon reduction of a series of three 4,4’bipyridine 3,3’-disulfide compounds has been investigated by means of cyclic
voltammetry and DFT calculations. These complexes contain two possible redox active
sites; the disulfide bond and the bipyridine/bipyridinium ring. The three compounds show
distinct cyclic voltammograms. The nonmethylated compound goes through an
irreversible two-electron reduction followed by a two electron oxidation. The
monomethylated compound undergoes two quasi-reversible one-electron reductions,
while the dimethylated compound undergoes four reversible one-electron reductions.
DFT calculations show that both reductions of the nonmethylated compound are centered
around the disulfide bond with S-S cleavage occurring after insertion of the second
electron. The LUMO of the monomethylated and dimethylated compounds is centered on
the bipyridine rings, and consequently that is where initial reduction occurs. Evidence of
intramolecular electron transfer is observed after the first reduction of the
monomethylated and the second reduction of the dimethylated compound. The
experimental cyclic voltammograms can be simulated in agreement with the proposed
DFT mechanisms.
Non-innocent ligands show rich and complex redox behavior which allow for
modulating the redox potentials of catalytic 2Fe2S centers; understanding the interaction
16
between the two redox centers allows for better design of catalytic systems, and is a
promising approach.
17
Chapter 1 Introduction
Natural Hydrogen Production
By employing evolution Nature has an incredible ability to use abundant elements
to efficiently catalyze seemingly simple chemical reactions which are in reality complex
and difficult. One example is the [FeFe]-hydrogenase class of enzymes which uses an
2Fe2S core to catalyze the production of molecular hydrogen from aqueous sources at
near thermodynamic potentials.1-3 Iron and sulfur are Earth-abundant elements, and thus
they are readily available and inexpensive. This is in contrast to current catalysts utilized
for the same conversion, such as platinum, which are rare and expensive. Consequently
use of complexes similar to that found in the active site is extremely attractive, and has
garnered much interest.4-10
The function of an enzyme active site is greatly impaired when removed from the
scaffolding of the enzyme itself, and synthetic preparation of an exact replicate of an
enzyme active site is often difficult, thus mass production of replicated active sites for
chemical goods synthesis is not practical. However, the demonstration of the ability of
2Fe2S clusters to catalyze hydrogen production does provide a starting point for mankind
to work from. Thus the main focus of this dissertation is studying the electronic
characteristics of compounds containing a 2Fe2S core for the purpose of designing better
hydrogen producing catalysts. This is accomplished by studying a series of benzo and
naphthoquinone substituted 2Fe2S systems, compounds where the 2Fe2S core has been
substituted with 2-phenylazopyridine, and the redox behavior of a series of 4,4’bipyridine-3,3’disulfide compounds via cyclic voltammetry, DFT calculations, and
spectroscopic techniques.
18
Hydrogenase Inspired Complexes
General Background
Anaerobic microorganisms are known to utilize the half reaction in scheme 1.2
for the storage of energy in the bond between two hydrogen atoms. There are three
distinct phylogenetic classes of hydrogenase enzymes, [NiFe], [FeFe], and Hmd or [Fe]only. While all three are active in both directions of the half reaction shown in scheme
1.2, the [FeFe] variety is the most active towards hydrogen production, and is capable of
producing molecular hydrogen in near neutral aqueous conditions at a rate of 6,000-9,000
s-1. 11
2H+ + 2e- ⇌ H2
(1.2)
Unfortunately, [FeFe]-hydrogenases are not thermally stable, and degrade in aerobic
conditions, making direct use of the organism for catalysis impractical. Consequently the
development of robust functional mimics for inexpensive and efficient production of
molecular hydrogen is needed. The active site of [FeFe]-hydrogenase can be seen in
Figure 1.1.
The study of [FeFe]-hydrogenase mimics falls into one of two broad categories;
biomimetics of the active site are used for a better understanding of the structural and
mechanistic behavior of the enzyme itself, while functional models aim to make strides
towards readily synthesized small molecule catalysts capable of producing molecular
hydrogen by economical means on an industrial scale. This dissertation focuses on the
latter, and thus examines the function and electronic structure of molecules which were
inspired by [FeFe]-hydrogenase, but deviate structurally in an attempt to increase
stability, synthetic accessibility, and production of molecular hydrogen.
19
The active site consists of an 2Fe2S butterfly core appended with a bridging
dithiolate, a cysteinyl ligand that bridges to an Fe4S4 cluster thought to act as an electron
transport chain, an unidentified ligand which could possibly be a substrate docking site,
but is generally presumed to be occupied by a water molecule, and a mixture of CO
and -CN ligands, both of which are electron withdrawing, uncommon in nature, and
generally toxic to living organisms. The prevalence of typically toxic CO and –CN
ligands suggests that their electron withdrawing ability due to π-backbonding is vital to
the function of the enzyme. Variations of 2 with a mixture of CO and –CN ligands and a
bridging linker as seen in 1 have been previously synthesized,12 and synthesis of nearly
exact models of the active site including attachment of the 4Fe4S H-cluster have been
accomplished,12,13but in both instances fail to be as catalytically active as the more
synthetically accessible Fe2(SR2)(CO)6 variants, such as 3 and 4.
The variation of ligands (L) attached to 2 in attempts to tune the electrochemical
behavior in favor of molecular hydrogen production is immense. The predominant
identity of L is CO, with most molecular examples containing 4-6 CO ligands with the
remaining sites occupied by ligands which have readily tunable electron donating
abilities such as phosphines or N-heterocyclic carbenes. As can be seen in Figure 1.1 the
active site contains a mere three CO ligands and two –CN ligands. The –CN ligands are
not as strong of π-backbonding ligands as CO, and thus with the greater number of CO
ligands in many of these mimics it is necessary to add electron density back. Phosphines
allow this due to strong σ-donation, and only slight π-backbonding by comparison to CO.
Additionally the R groups on phosphine ligands can be varied greatly, tuning the σ-
20
donation and π-acceptance of the phosphine, and thus the electron density at the metal
centers. The behavior of N-heterocyclic carbenes is analogous to that of the phosphines.
Additionally, the identity of R in complex 2 varies from “open systems” in which
the R groups are not linked to one another to systems such as 3 with functionalized
versions of the saturated linker seen in 1, to ene-dithiolates as seen in 4. Recently,
attention given to attachment of chromophores for capturing the energy of light to drive
catalysis,14 and surfactant type ligands to increase solubility in water has increased. Due
to the wide range of complexes being examined in the literature, and the existence of
several hundred variations of 2, the review of literature herein will focus on two basic
structural motifs, 3a-c and 4, that serve as a primary starting points for understanding the
electronic structure, and redox behavior of this class of chemicals, and some variations
thereof that strive towards photoproduction of molecular hydrogen. Many of the variants
of 2 are either synthesized directly from 3a-c and 4, or they share considerable structural
similarities.
21
Figure 1.1 The [FeFe]-hydrogenase active site, shown with synthetic variants inspired by
[FeFe]-hydrogenase.
22
In acetonitrile complex 4 undergoes a reversible two-electron reduction at -1.32 V
Vs. Fc+/Fc.15 Agreement between calculated structures, calculated reduction potentials
and values obtained via simulation of the electrochemical parameters suggest that the
potential inversion is due to the lengthening/breaking of the Fe-Fe bond, breaking of an
Fe-S bond and rotation of a CO ligand into a semi-bridging structure. The calculated
structures of the dianion have recently been substantiated by determination of the single
crystal X-ray structure of the chemically reduced dianion.16
Complex 4 produces molecular hydrogen at a potential of -2.1 V Vs. Fc+/Fc in
acetonitrile with acetic acid as the proton source. With a thermodynamic potential for the
conversion of acetic acid to molecular hydrogen of -1.46 V Vs. Fc+/Fc, this gives an
overpotential of -0.57V. The electrochemical mechanism for this has been investigated
and it was determined that the dianion is protonated, followed by reduction, then
introduction of a second proton forms molecular hydrogen and produces the monoanion,
which is immediately reduced to the dianion at the potential at which catalysis is
occurring. Thus, if sufficient acid is present the monoanion is able to immediately reenter
the catalytic cycle without returning to the neutral species. Thus, complex 4 is considered
to be a procatalyst, while the first active catalytic species is 4- which cannot be isolated
due to potential inversion.
Complex 3a has been shown to have a very similar solid state structure to that of
1 as present in the enzyme.17 In acetonitrile at 100 mV/s 3a shows an oxidation peak at
0.74 V, and a quasi-reversible reduction peak at -1.65 V Vs. Fc+/Fc. The lack of
reversibility in the peak suggests possible degradation of the compound, and while the
decomposition product could not be isolated from the electrochemical solution, a dimeric
23
structure was proposed to be the decomposition product based on spectroscopic data.18
Indeed a dimer structure has been synthesized chemically19 and is believed to be the
cause of the loss of anodic current on the return wave of the cyclic voltammogram. For
this reason redox events occurring negative of the first reduction cannot be assigned
exclusively to 3a-, and consequently redox events negative of the first reduction will not
be examined here for 3a-c. Production of molecular hydrogen from acetic acid occurs
at -2.35 Vs. Fc+/Fc giving an overpotential of 0.89 V. Additionally the catalytic current
has been reduced by comparison to 4. The generally accepted mechanism for 3a with
weak acids is an ECEC mechanism.
Of the often studied complexes, 3b most closely resembles 1 in structure with an
identical bridge between the two sulfur atoms. 3b, is surprisingly less studied than 3a.
The original synthesis of 3b was reported by Rauchfuss et al.,20,21 and preliminary
electrochemical studies performed by Sun and coworkers
20
show 3b to have an
irreversible oxidation at 0.59 V and an irreversible one-electron reduction at -1.58 V Vs.
Fc+/Fc in CH3CN. Attempts made by Sun et al. to protonate 3b with triflic acid, were
unsuccessful, though only the neutral complex was examined. Currently, the literature
has not examined the ability of 3b to catalyze the electrochemical reduction of acetic acid
to molecular hydrogen. Similar to 3a, 3b has been functionalized2220,23-30 in a variety of
manners with much of the work focusing on replacement of carbonyl ligands with
phosphines and N-heterocyclic carbenes, or substitution of the amino hydrogen atom.
Complex 3c was originally synthesized by Rauchfuss et al. in 2002,21 but
electrocatalytic hydrogen generation was not explored until 2005 when Song et al.
investigated the redox behavior in acetonitrile.31 3c exhibits an irreversible oxidation at
24
0.81 V, and an irreversible reduction at -1.59 V Vs. Fc+/Fc. Catalytic production of
hydrogen from increasing concentrations of acetic acid occurs between roughly -2.0 V
and -2.1 V, giving an overpotential comparable to that of 4, but with diminished catalytic
current.
Examination of the oxidation potentials of 3a-c shows no clear correlation with
the electronegativity of X. 3a and 3c have approximately the same oxidation potential,
while 3b shifts negatively by slightly over 0.1 V. The reported reduction potentials of 3b,
and 3c are nearly identical while the reduction potential of 3a is shifted negatively by
approximately 0.06 V. As is the case with the oxidation, the literature does not provide a
clear correlation between the electronegativity of X and the potential of reduction.
Implications of Non-innocence
The [FeFe]-hydrogenase active site relies upon the structural environment of the
protein and an electron transport chain capable of long range electron transfers via
variable potential gradients in order to achieve the minimal overpotential at which it
operates. While both of these techniques have been mastered by Nature on the
evolutionary time scale; the timescale of the life of a chemist is considerably shorter, and
thus such a complicated system is not easily achievable. Due to this unfortunate
constraint of a finite human lifetime it becomes apparent that the astute chemist will have
to be clever in order to match or best Nature. For this reason the interaction of 2 with
ligands which participate in the redox chemistry of the complex, referred to henceforth as
non-innocent ligands, and ligands which strongly absorb visible light, or chromophores
are being examined.
Non-innocent ligands came to prominence in the 1960’s and have been maturing
and coming to age since that time.
32
The unique ability of non-innocent ligands could
25
also lend themselves to facilitating H2 production by being reduced in a more facile
manner, and delocalizing the electron to the diiron core where catalysis occurs. In this
manner the overall complex will have been reduced at a less negative potential, and
ideally be able to perform catalysis closer to the thermodynamic potential.
Ene-dithiolate ligands such as 1,2-benzenedithiolate as employed in complex 4
are a common example of non-innocent ligands. Enemark and coworkers have used 1,2benzenedithiolate ligated to transition metals extensively for the purpose of modeling the
active site of sulfite oxidase enzymes which contain a Mo active site ligated with a
pyranopterin dithiolate.32-39. It is known that the dithiolate coordination is able to
facilitate conversion between redox states of the enzyme active site by changing the
overlap of the S pπ with the metal center. Employing such concepts to mediate the redox
potential of [FeFe]-hydrogenase mimics could prove beneficial to catalysis.
This idea is directly employed in chapter 3 where the electronic communication of
the 2Fe2S core of a series of enedithiolates, (μ–dithiolato)Fe2(CO)6 complexes, in which
the sulfur atoms are linked by 5–substituted–1,4–benzoquinones(Me, OMe, Cl, t–Bu),
1,4–naphthoquinone or 1,4-anthraquinone is studied. The redox behavior of these
complexes in solution is examined and compared to the free quinones. The shift in redox
potential from the free quinone to the (μ–quinonedithiolato)Fe2(CO)6 show the diiron
hexacarbonyl moiety to have approximately the same electron withdrawing ability as
substitution of quinones by three chlorine atoms. DFT computations and EPR
measurements are used to examine communication between the quinone moiety in the
once reduced species; it is shown that substitution effects at the 3 and 4 positions of the
quinone greatly affects radical character on the metal centers.
26
Similarly, the reduction behavior of a series of methylated 4,4’-bipyridine-3,3’disulfide complexes are examined in chapter 5 as potential disulfide ligands for these
systems. Methylation of the N atoms on the bipyridine rings changes the reduction
behavior.
Utilizing Non-Innocent Chromophores to Harvest Light.
Chromophores are molecules which strongly absorb light, and when ligating a metal
center can potentially behave non-innocently. Ideally when attached to the diiron core of
2, energy captured by the excitation of electrons by visible light will facilitate electron
transfer from the excited state chromophore to the diiron core. The vacated hole of the
chromophore could then be reduced at a much less negative potential than the ground
state of either molecular fragment. In theory, with the aid of sunlight, production of
molecular hydrogen can be achieved at an underpotential, i.e. with less energy than the
thermodynamic potential being supplied from an electrode. This basic principle is
presented in Figure 1.2. A variation on this principle would be for the chromophore and
diiron centers to not be covalently linked. This offers the advantage of inhibiting reverse
electron transfer, but comes at the cost of relying upon the excited chromophore to
collide with a variation of 2 during the lifetime of the excited state in order to allow for
an
intermolecular
charge
transfer.
27
Figure 1.2. (a) ground state of a neutral catalyst with attached chromophore. (b) excited
state of a neutral catalyst that is in the process of being reduced at a lower energy due to
electronic excitation. (c) reduced catalyst
28
Figure 1.3 Two examples of [Fe-Fe]-hydrogenase inspired molecules designed for light
harvesting.
29
The first example of a variant of complex 2 being ligated by a chromophore
occurred in 2003 in an attempt by Sun and Åkermark.40 The complex synthesized, 5, was
a variation of 3a in which the central carbon of the propane bridge was substituted with a
ruthenium bipyridine complex connected via a p-amidobenzoate linker (Figure 1.3). The
UV-Vis and redox behavior of complex 5 is dominated by the Rubipy portion of the
molecule with minimal change from that of the free Rubipy, suggesting minimal
electronic communication between the Fe and Ru metal centers. This is not surprising as
the Ru and Fe centers are separated by a large distance, and lack π-conjugation between
the two. In a subsequent papers41 in continuation of the study by the same research
groups, a substituted version of 3b, in which the N-H had been replaced by an N-benzyl,
was studied in the presence of free Ru(bipy)32+ in solution. It was determined that the
redox potential of the excited state of Ru(bipy)32+, Ru(bipy)33+/2+*, was too positive to
reduce the FeIFeI catalyst. However, if ascorbic acid was used as a sacrificial electron
donor and acid, the in situ generated Ru(bipy)31+, had a sufficiently negative redox
potential to reduce the catalyst of interest. This three component system was shown to
achieve a turnover number of 4.3 based on the Fe catalyst.
Sun et al. made improvements on this system by synthesizing 6, where a zinc
tetraphenylporphyrin (ZnTPP) complex was reversibly linked to a hydrogenase mimic via
axial coordination to a N atom, as seen in Figure 1.3. The idea being that after the ZnTPP
complex is excited and an electron is transferred, the linkage between pyridine and
ZnTPP will break, thus back electron transfer will be prevented and the catalyst will
effectively be reduced at the potential required to backfill the ZnTPP HOMO.42 This
30
worked well as a proof of concept and exhibited photoinduced H2 evolution, albeit with a
turnover number (TON) of 0.16, making it unclear whether this complex should be
considered a catalyst or reactant. Some advances have also been made in using noncoordinated nanoparticles with water soluble variations of 3a to produce a TON of
around 500.14
Chapter 4 focuses on the study of two complexes in which two carbonyl ligands
on
1,3-(µ-propanedithiolato)diironhexacarbonyl
1,2-(µ-benzenedithiolato)diironhexacabonyl
(4)
have
(3a)
been
substituted
and
by
the
chromophore 2-phenylazopyridine (PAP). Phenylazopyridine, 3a and 4 are orange-red in
color but their PAP derivatives are an intense blue color, suggesting charge transfer;
however, they do not show a dependence of λmax on solvent polarity. This discrepancy
and the nature of PAP bonding to 3a and 4 is examined with UV-Vis, and IR
spectroscopies, as well as TD-DFT.
Cyclic Voltammetry
Cyclic Voltammetry is an electrochemical technique for the study of solution
redox potentials of a molecule. The standard setup employed in our research lab consisted
of three electrodes, a working electrode, counter electrode, and reference electrode. The
potential difference between the working electrode and reference electrode is varied
linearly with time while the current of the working electrode is monitored to indicate
redox processes occurring at the surface of the working electrode. The counter, or
auxiliary electrode, serves to complete the circuit and thus behaves as the anode when the
working electrode is acting as a cathode and vice versa.
The results obtained from a cyclic voltammetric experiment are displayed as the
measured current vs. the applied voltage. An example of a two-segment cyclic
31
voltammogram for ferrocene can be seen in Figure 1.4. Much information can be gleaned
from examination of a voltammogram. For example, in Figure 1.4 one can see that the
anodic peak and cathodic peak area beneath the curve, have nearly the same peak current,
and are separated by approximately 72 mV. This is close to the theoretical separation of
59 mV for a solution-phase electrochemically reversible process at 25 °C, and the
deviation from 59 mV in this example is due to uncompensated solution resistance.
Assuming a similar peak shape on the anodic and cathodic waves, one can estimate
reversibility by taking Ipa/Ipc. Ignoring diffusion this would give a theoretical value of 1.0,
but due to diffusion in solution the value for a fully reversible process will always be
somewhat less than 1.0 and in the case of Figure 1.4 gives a value of 0.96. This shows
that ferrocene is reversible on the electrochemical time scale.
Reversibility is important for catalysts as it suggests that they may be more robust
than examples which are not reversible. It is important to note that what is commonly
referred to as chemical reversibility is not necessarily the same as electrochemical
reversibility. Comparison to the Fc+/Fc couple is a good baseline to measure the
reversibility of catalysts. In addition to probing the redox behavior of the catalyst itself,
the redox properties are also studied in solutions with various concentrations of acid,
allowing for investigation of the reaction shown below.
2HA + 2e- ⇌ H2 + 2AAs the concentration of acid is increased, an increasingly large catalytic current
should be present corresponding to a peak for the catalytic production of molecular
hydrogen and a conjugate base. The catalytic current and width of the peak will allow
information to be gained about the amount of hydrogen produced on the timescale of the
32
experiment, and the potential the peak occurs at will allow for determination of the
excess energy put into the system beyond the thermodynamic potential. The difference
between the potential that catalysis actually occurs at and the thermodynamic potential is
referred to as the overpotential, and will be referred to as such for the remainder of this
dissertation.
33
Figure 1.4 Cyclic voltammogram of a dichloromethane solution of 1.08 mM ferrocene,
and 0.2 M tetrabutylammonium hexafluorophosphate on a glassy carbon electrode
showing both empirical data (green line) and simulated voltammogram (black circles).
34
Density Functional Theory Calculations
Density Functional Theory calculations have increasingly become vital to
understanding mechanisms and electronic structure in chemistry.43 The complexity of the
molecules studied herein makes ab initio calculations impractical due to the immense
computing power and time required for systems of this size.43Additionally DFT has been
proven to be sufficiently accurate for this class of complexes,44,45 and will be employed
within this dissertation. For many of the complexes pertaining to this body of work
functionals and basis sets have been extensively tested and shown to correctly predict
nuclear magnetic constants and spin states for iron complexes.44,45 Consequently DFT
calculations allow for the modeling of the electronic structure of these complexes to a
degree which would otherwise not be easily obtainable.
35
Chapter 2 Experimental Methodologies
Electrochemistry
For electrochemical experiments, the source and treatment of the solvent and supporting
electrolyte have been described earlier.46 Electrochemical data was collected in
acetonitrile or dichloromethane with various concentrations of tetrabutylammonium
hexafluorophosphate (Bu4NPF6) as supporting electrolyte, as laid out in the specific
chapters. The potentiostats employed during this dissertation varied between an EG&G
PAR model 273, a BAS CV-50W, and a Gamry Instruments Reference 3000. For cyclic
voltammetry the working electrode was a 3 mm glassy carbon disk electrode (0.088 cm2)
or a mercury–film on a gold–disk electrode (0.080 cm2 prepared as described earlier).47
The reference electrode was made of a silver wire in contact with a solution of 0.01M
AgNO3 in solution with the same solvent and electrolyte concentration as the analyte
being studied. At the end of electrochemical experiments the Fc+/Fc couple was
measured and the data was shifted to set the Fc+/Fc couple at zero potential. In most
instances electrolyte solutions were scanned over the same potential window at the same
scan rate to subtract charging current. For voltammograms performed on the EG&G,
solution resistance was ascertained by studies of the oxidation of ferrocene, whose
diffusion coefficient is known.46 Evaluation of solution resistance was carried out as
described earlier,46 and the resistance was partially compensated by electronic resistance
compensation with the remainder of the resistance applied when simulating the data.
When using the Gamry Reference 3000, solution resistance was measured at 0 V relative
to AgNO3 with the instrument in conjunction with the supplied software, and 80% of this
value was compensated for with positive feedback. Voltammetric experiments were
carried out at room temperature.
36
Digital simulations were conducted with DigiElch, version 7.0, a software package for
the
Digital
simulation
of
common
Electrochemical
experiments
(http://www.digielch.de).48 The fitting routine in that program was used to establish the
final best–fit parameter values for many of the variables.
Notes to Future Team DLL Students on Performing Cyclic Voltammetry
Below is a little more specific instruction for Team DLL students picking up where I left
off on the way that I ran electrochemical experiments on our new Gamry Ref 3000. The
Ref 3000 is a powerful potentiostat, capable of performing more experiments than
covered here. Most, but not all of the software provided by Gamry for this potentiostat is
open source. Thus, if you do not like the software interface described herein you can
modify the display windows to give you more or fewer options. Gamry charges for
software packages in addition to the instrument itself. The instrument was purchased with
only the pulse voltammetry package and the physical electrochemistry package. If the
experiment you wish to run is not included in these two software packages, then there is
still a strong possibility you can order the correct software package and still use this
instrument.
Preparing Solutions
Most of the compounds we perform cyclic voltammetry on are run at a
concentration of 1 mM analyte. If little compound is available 0.5 mM analyte can easily
be used. On past instruments it was difficult to go much below 0.5 mM, but with the
sensitivity of the new Gamry instrument it may be possible to use a lower concentration.
Typically a concentration of 0.1 M tetra n-butylammonium hexafluorophosphate (TBAH)
is used as supporting electrolyte for measurements performed in CH3CN, and 0.2 M for
measurements made in DCM or DMF. With concentrations above 0.3 M electrolyte
37
problems with ion-pairing are encountered. Owing to the high cost of the electrolyte,
~$2/g, the used solutions are emptied into a waste beaker and the solvent is allowed to
evaporate so the electrolyte can be recovered. The easiest method to recrystallize the
electrolyte is to first dissolve the contents of the waste beaker in acetone (TBAH is
extremely soluble in acetone) and filter to remove any rust or other insoluble impurities.
From here perform a series of recrystallizations in EtOH until pure.
Preparing the Cell and Electrodes
The cell used for cyclic voltammetric studies can be seen in Figure 2.1, it is
designed around a volume of 10 mL, but can easily be used with about 5 mL. If anything
happens to the cell we have a backup made, but you will have to take it to the glass shop
to have the electrodes/cell necks bent so they will all fit in the cell together. After
extended use the cell can start to have a slight orange color to it, if this happens clean
with aqua regia.
We use an in-house reference electrode that is not commercially available. If the
glass body of the electrode breaks, the glass shop can repair it or make you a new one.
There are actually two compartments to the electrode. The top is the actual reference
electrode,
38
Figure 2.1 Electrochemical cell for cyclic voltammetry employed by our laboratory
showing A) Pt counter electrode with air vent. B) Glassy carbon working electrode. C)
Inert air bubbler. D) AgNO3/Ag counter electrode with salt bridge. Below shows
electrode connectivity.
39
while the bottom is a salt bridge. When running in CH3CN the solution in the upper
compartment of the reference electrode is 0.1M TBAH, and 0.01M AgNO3 in CH3CN in
contact with a silver wire. The lower compartment is 0.1M TBAH in CH3CN. AgNO3
photodegrades, and consequently the solution in the upper cell needs to be replaced
approximately once a month. You can keep a stock solution of AgNO3in a dark chemical
cabinet to refresh the electrode.
The tips of the two compartments are made of Vycor, they cost ~$80 for a 3-pack
through BASi, but could be obtained from other sources as well. In order to preserve the
tips and keep them from clogging with salts, they need to be stored in electrolyte
solution. Do not allow the electrodes to be stored dry. When they do eventually
become clogged you will start to notice additional noise in your voltammograms.
Typically if the electrode tip is well cared for it needs to be replaced every 1-1.5 years.
The Vycor is held in place with PTFE/Teflon heat shrink tubing. As of right now there is
~3 feet of 8 gauge 0.008” wall PTFE heat shrink tubing in the space below the IR. In a
pinch you can use regular heat shrink tubing from the electronics shop in CH3CN, but it
immediately dissolves in DCM.
When using the glassy carbon electrode, you should polish the working electrode
on a felt pad in conjunction with 0.05 micron alumina. I usually polish for 3-5 minutes.
The duration of polishing depends on how hard you are pushing down on the electrode.
When polishing it is important to keep the electrode flat against the felt pad so that you
are not over-polishing one side and under-polishing the other. A quick search of the web
will give numerous manufacturer pages with detailed instructions for electrode polishing.
After polishing, rinse the electrode with deionized H2O to get rid of any alumina on the
40
surface, and then flick it with my wrist to get rid of as much water as possible. Any water
remaining on the shank of the electrode gets wiped off with a kimwipe, and if any is left
on the face of the electrode wick it away with a kimwipe without touching the surface. At
the end of the days experiment you should also make sure to wash the felt pad free of
alumina so that it does not aggregate.
The counter electrode is made of Pt, and requires little maintenance. Rinse it with
acetone at the end of the experiment, and if it takes on to much color you can dip it in
acid and rinse it off, but that is rarely necessary.
If you wish to run under an inert atmosphere, now is the point to start sparging the
solution with Ar while stirring. Note that there is no need to run under an inert
atmosphere if you are studying Fc, and it would be counterproductive to do so if you are
studying O2. The flow rate I use normally rids the solution of O2 in about 20 minutes. I do
not like to sparge at a faster flow rate so as to minimize solution evaporation. While
sparging connect the electrodes as seen in Figure 2.1. The white electrical lead is to be
connected to the reference electrode, red is to be connected to the counter electrode, and
both the green and light blue are to be connected to the working electrode. The blue
electrical lead is the sensing working electrode. Basically it has a high impedance resistor
in its circuit so uses little current but is able to accurately measure the potential of the
current coming off of the working electrode. Additionally there is a lead coming from the
rear of the instrument which is an earth ground and should be connected to a water pipe
or some other large piece of metal. You will be left with two leads that are not connected.
The black is a floating ground electrode that would be connected to a faraday cage if you
were using one, and the orange lead is a sensing lead for the counter electrode. The
41
orange lead is used for Zero Resistance Ammeter (ZRA) mode, in which the working and
counter electrode are held at the same potential and the current flow between the two is
measured. Because we do not control, or really even care about the potential at the
counter electrode in cyclic voltammetry, this lead is not used.
The potentiostat itself is controlled with the Gamry Framework program. Note
that you need to have the potentiostat on prior to initializing the program or when you go
to do anything you will get an error that reads “Pstat device list is empty. [Common
Functions.exp (line 197)]” If this shows up just shut down the program, turn on the
potentiostat, and initialize the program again.
For cyclic voltammetry you do not want an agitated solution, so at this point if
you have your solution sufficiently purged and everything set up, switch the flow of Ar
from through the solution to above the solution, and turn the stir plate off via the surge
protector mounted to the right-hand side of the cart. If power to the stir plate is not
removed oscillation in the current will be present due to RF from having the stir plate
plugged in (shows the sensitivity of the potentiostat).
Estimating Solution Resistance
The first experiment is to measure/estimate the solution resistance. This is not
100% accurate, but works pretty well for estimating the amount of solution resistance we
can compensate for with the instrument, and not lose control of the potentiostat. The
program is accessed through the experiments drop down (Figure 2.2A). There is a 99%
chance this will be one of the last 8 experiments performed and you can click on the
program c:\programdata\gamry instruments\framework\getru.exp under the most recently
used list. If it is not one of the last 8 experiments you can access it through the Named
Scripts option, or through A Utilities. Note that any program that is from Gamry directly
42
is available in A Utilities, B PHE 200 – Physical Electrochemistry, or C PV220 – Pulse
Voltammetry, if the program has had the code modified by someone in the lab it must
either be accessed through the most recent programs or Named Scripts. The instrument
should estimate the solution resistance of 0.1 M TBAH in acetonitrile at approximately
150 Ω. Multiply whatever the number is by 0.75 and this is the amount of solution
resistance you will compensate for with the instrument.
At this point you need to decide how many segments you would like to have in
your cyclic voltammogram. Figure 2.2B shows the way a cyclic voltammogram is run,
and how many segments you will wish to perform. A linear sweep voltammogram would
be sweeping from point t0 to t1, a 2-segment would follow the entire black trace, while a
3-segment scan would follow the black trace and continue through the red trace. If you
are beginning your experiments by performing experiments on Fc, choose to run a 2segment experiment, if you start with O2, it will be the same procedure, but with a
different window and initial direction. A 3-segment scan could be used, but is slightly
more cumbersome when working up data in Excel.
Performing a 2-segment Scan
To perform the 2-segment scan, go to the Experiments drop down, and either
select :\programdata\gamry instruments\framework\2-segment CV.exp from the recent
experiments list, or navigate to it through Named Scripts. The window that appears is
displayed in Figure 2.3. The “Test Identifier” and “Notes…” are for your own personal
use and to be used as you see fit. When you name your output file it is best include the
43
Figure 2.2 A) Experiments tab for Gamry Framework. B) Graph of a generic cyclic
voltammogram displaying Potential Vs. Time
44
Figure 2.3 Gamry Framework display window and settings used to perform cyclic
voltammetry on 1 mM Fc in acetonitrile.
45
.dta extension. If you forget to include this extension, the file will still be written and it
will still include all of your data, but the extension will have to be manually added if you
want to use other Gamry programs to open it. Initial E is the potential you wish to start
your experiment at vs the AgNO3/Ag reference electrode (E at t0 in Figure 2.3). Scan
Limit 1, also referred to as the switching potential or the vertex 1 on some systems
corresponds to where you will reverse scan direction and is listed as v1 and t1 in Figure
2.3. Generally the Final Potential is left to be the same value as the initial potential but
does not necessarily need to be the case. For the study of the Fc+/Fc couple the Final E
would correspond to t2 on Figure 2.3, with the entire cyclic voltammogram being
represented by the black portion of Figure 2.3. The values for Fc+/Fc Vs. AgNO3/Ag,
depending on exact concentrations of AgNO3 and the age of the electrode, will be
approximately:
Initial E (V)
-0.3
Scan Limit 1
0.5
Final E (V)
-0.3
You may set the Scan Rate at the range you like. The used default is 100 mV/s,
and if you go much below 50 mV/s you begin to encounter problems due to too much
diffusion towards and away from the electrode.
The Step Size (mV) determines the frequency of your sampling, and when setting
up the potentiostat a value of 2 seemed to work well over a wide range of scan rates
without creating oscillations due to oversampling.
You want the IRComp to be set to Positive Feedback (PF) and under PF Corr.
(ohm) enter 75% of the value you obtained from the getRu.exp experiment.
46
Turn Conditioning on, and set to the potential where your scan begins. Fifteen
seconds is sufficient, but this value may be increased. This function holds the potential
difference between the working electrode and the reference electrode at your specific
value, for your specified time. Any charge at the tip of the working electrode, will affect
the ordering of the ions in solution around the tip. As the ions reorganize additional
charge will build on the tip of the electrode and cause more ion reorganization, thus
creating the electrical double layer. This mathematically can be thought of as a capacitor,
and thus as it forms current is passed in order to charge the capacitor. By enabling this
function we are removing this charging current from the beginning of the cyclic
voltammogram.
Depending on what experiment you are performing, you may wish to have the
advanced Pstat Setup ticked on. What most frequently needs changed in this window is
the filtering options. Under a scan rate of 3 V/s you should set both filters to 1 kHz,
above 3 V/s you should set it to the 200 kHz filter. Limited testing with the filters on
auto, showed the program generally set them in the same as suggested above, but it is
preferable to set them manually, rather than using the auto settings. Clicking OK here
will run the experiment. Watch the data appear on the screen to ensure that the instrument
does not go into overload. If an overload does occur, press F1 to abort, and Y for yes
when it asks if you are sure. If an overload occurs, go back through this and make sure
you followed directions and that all electrical connections are secure. Also check your
solution resistance and make sure you are not compensating for too much. If you still
have a problem check to see if you have any air bubbles around your electrode tips. The
settings displayed in these figures should allow for collection of a CV for ~ 1 mM Fc, and
47
0.1 M TBAH in CH3CN with a 0.01 M AgNO3/Ag reference electrode. Assuming
everything went well with your first CV, stir the solution for 5-10 seconds, let sit for ~15
seconds and run another. Do this so that you have three cyclic voltammograms all under
the same conditions.
Once you have three cyclic voltammograms, open Gamry Analyst and direct it to
all three of your files (they automatically save to the folder My Gamry Data, a shortcut is
on the computer’s desktop). For well-behaved compounds these three CV’s should lay
more or less perfectly on top of one another. Sometimes the first differs slightly because
the conditioning the electrode is slightly different between the first and second scan, or
you have some residual alumina on the surface which is removed after the first scan. As
long as the second and third scans are consistent this is satisfactory. There are additional
features of Gamry Analyst that can be useful such as peak picking, integration,
background subtraction, and current normalization.
3-segments Scans
If you are studying one of the group’s catalysts, you will come to a point where
you wish to display the compounds reductions and oxidations on the same
voltammogram. In order to do this you will need to perform a 3-segment scan, which is a
combination of the black and red traces in Figure 2.2B. When doing this, use the program
c:\programdata\gamry
instruments\framework\cyclic
voltammetry.exp.
There
are
modified versions of this script under slightly different names with additional options
turned on and other features turned off (such as fixing the electrode area to 0.0878 cm 2)
so they can’t be changed. The procedure for this script is basically the same, but now you
have access to Cycles # (number of times the cyclic voltammogram repeats) and a second
48
scan limit (v2 in Figure 2.2B). A typical scan for one of our Fe2S2(CO)6 compounds
might be:
Initial E
0V
Scan Limit 1
-2.5V
Scan Limit 2
1.3 V
Final E
0V
If you wish to incorporate other features such as a time delay at one of the vertices, the
script will need to be modified.
Fc+/Fc Correction and Backgrounds
Once all of the voltammograms on your compound have been collected, you will
need to run voltammograms on Fc+/Fc and on a solution of just solvent and electrolyte
for use with the ferrocene correction and background subtraction. If the compound does
not show any redox process in the same area as the Fc+/Fc couple then add a small
spatula tip (1-5 mg) of Fc to the solution and allow it to dissolve; then run three scans
stirring between each scan. Overlay the files and find the min/max in Gamry Analyst.
The Fc+/Fc couple should not vary more than 1-2 mV between all of the scans.
When done with your solution, pour the solution into the waste beaker to
recapture the electrolyte. Wash the cell and all electrodes with the acetone. and leave the
cell upside-down to dry on chemisorbent pads. While that is drying measure out
electrolyte for a background solution. You want to run backgrounds for every
voltammogram that you will try to model in DigiElch, or present to others. The
background should be run identical to the voltammograms you obtained (i.e. use the same
step size, window, scan rate, filter, ect.).
49
Data Work-Up
For the data work up you need to accomplish 3-4 things: reference your potentials
so that the Fc+/Fc couple is set to zero, subtract the background current from your current
for each voltammogram, if performing a scan rate study normalize the currents by
dividing by the square root of the scan rate (in V/s) and graph the voltammograms.
Follow the American convention of plotting CV’s, Gamry displays IUPAC convention,
and also multiply the current so they y-axis is in μA instead of A. A macro could be
written to accomplish many of these, and it is advisable to do so if you will be performing
a lot of electrochemistry throughout your graduate career.
EPR Spectroscopy
All EPR spectra were collected on species that were generated electrochemically
in
situ
via
bulk
electrolysis
in
DCM
with
0.5
M
tetrabutylammonium
hexaflourophosphate as a supporting electrolyte. The spectroelectrochemical cell is made
of quartz with a platinum counter and working electrode, and a Ag/AgNO3 reference
electrode. An X–band Bruker Elexsys E500 spectrometer was used for all spectra.
Spectra were modeled with EasySpin software49 implemented into MATLAB® to extract
hyperfine constants and g–values.
Computational Methods
All computations performed in this dissertation were carried out using
ADF2009.01,43,50 ADF2012.01,
51
ORCA52 or Gaussian09,53 as noted in the chapters.
Geometry optimizations and frequency calculations for inorganic compounds were
carried out using the VWN functional with the Stoll correction implemented.54 Figures of
the optimized geometries, and electron distribution were created with the program Visual
50
Molecular Dynamics 1.9.55 All EPR hyperfines reported were calculated using the OPBE
density functional.44 Recent comparisons of OPBE to other common functionals found it
to be the best for the prediction of nuclear magnetic constants45 and the only functional to
correctly predict the spin states of seven different iron complexes.44 Comparisons with
other common functionals in the ADF package have also shown it to be among the best at
predicting the oxidation and reduction potentials of several iron complexes and the pKa
values of the acids. A triple– STO basis set with one polarization function (TZP),
available in the ADF package, was used in all calculations. Relativistic effects were taken
into account by using the scalar ZORA formalism for geometry optimizations and spin
populations, and the spinorbit ZORA formalism for all hyperfine values,56 implemented
as part of the ADF program. All electronic structures with unpaired spin were calculated
using an unrestricted framework. Only low–spin complexes have been analyzed.
Carbonyl stretching frequencies were calculated at the conclusion of geometry
optimizations, and frequencies were examined for negative and nonreal frequencies. For
the purpose of comparing to solution (hexanes) spectra all energy values for a given
complex were shifted by a value of up to 1.006, though the exact value varies with the
molecule studied.
Infra-Red Spectorelectrochemistry
All IR spectroelectrochemical measurements were made in a thin layer reflectance
cell of similar design to what has already been reported in the literature.57 All
experiments employed an approximately 0.25 mm spacer. Thicker spacers allowed for a
longer path length, but it was found that diffusion also increased to the detriment of
electrochemical control. A CaF2 window was used in all experiments. Solutions were
51
made by dissolving 0.2 M TBAH and sparging for 20 minutes prior to addition of the
analyte. Background spectra of TBAH solution were collected and manually subtracted at
the conclusion of the experiment. All compounds studied in the spectroelectrochemical
cell had been previously studied by cyclic voltammetry. Cyclic voltammograms were run
for all compounds within the cell to determine potentials of redox events against the
pseudo silver electrode.
Notes to Future Team DLL Students on Performing IR-SEC
The cell that our lab has is directly based on a design out of the Kubiak group.57 The cited
paper explains its use very well. A CaF2 salt plate is used with the SEC cell rather than
NaCl or KCl, since it is more resistant to dissolution by semi-polar solvents. CaF2 has
limited solubility even in water so it can be used for aqueous IR-SEC. Also, in dry
climates, it does not need to be stored in the desiccator. The suggested spacer length of
0.25 mm to works well. A longer spacer length makes obtaining decent electrochemistry
difficult due to increased diffusion between the working and counter electrodes. The cell
does not require a large volume, so normally a solution of ~1 mg of analyte in 1 mL of a
0.3 M TBAH acetonitrile solution is used. Set the IR instrument to only take 1-2 scans
because the lifetime of species generated electrochemically may be relatively short. The
combination of a high concentration of acetonitrile and TBAH causes fairly strong
absorptions in the CO region, so run your background IR spectrum first. This also has the
added benefit of checking the cell for leaks before placing your compound in it. I
introduce the compound via a syringe. When I wish to purge the system of background
solution and add the compound I push solution through with the syringe, and wick the
overflow on the opposite side into a kimwipe.
52
Chapter 3 Redox Active Quinones Coupled to 2Fe2S Cores
Pertaining to Molecular Hydrogen Production
Part of this dissertation chapter have been accepted for publication in Organometallics58
Introduction:
As part of progress toward more efficient artificial systems for water splitting,
better catalysts for oxidation of water to evolve molecular oxygen and better catalysts for
reduction of the concomitant protons to evolve molecular hydrogen are being sought.
One broad class of hydrogen–evolving catalysts is inspired by the active sites of [FeFe]–
hydrogenases,1-3 which are extremely efficient enzymes for reversible reduction of
protons to hydrogen with exceptionally high turnover rates at very modest
overpotentials.59,60 The common feature of this class of catalysts is a butterfly 2Fe2S core
as depicted schematically in 2. In addition to the advantage of earth–abundant iron and
sulfur comprising the core, a variety of ligands can be bound to the iron centers and a
variety of functional moieties can be attached to the sulfur atoms to optimize the
reduction potentials, facilitate core and ligand geometry reorganizations that favor the
mechanism, shuttle protons to the reduction site, accelerate electron transfer, and couple
electronically for photoexcitation. According to a recent search of the CCDC61,62 more
than 600 catalysts of this class have been synthesized and characterized, and all catalyze
the reduction of protons to hydrogen under appropriate conditions.6 In most cases, H2
production occurs at potentials more negative than thermodynamic potential. Some
examples are known where the overpotentials are modest but so also are the rates for H2
production.4
A possible solution to obtaining a simple organometallic analogue of the 2Fe2S
active site that is an efficient electrocatalyst for H2 production with minimal wasted
53
energy is to append the 2Fe2S active site with an easily reduced organic moiety.
Furthermore, if the organic ligand is electronically coupled with the 2Fe2S catalytic
site,63 a complex with the desired activity might be obtained. There are pertinent
questions about which organic redox–active moiety to use and how to attach it to the
2Fe2S core to achieve correct orbital overlap, energies, and symmetry to assure electronic
coupling.64 Due to 1,4–benzoquinones being classic reversible, facilely reduced organic
compounds, they were chosen as the organic ligand to be annulated to the sulfur atoms of
the 2Fe2S site thereby bridging the two sulfur atoms as an approach to a system with the
desired properties. A series of such complexes were synthesized (Figure 3.1) and
structurally characterized by members of the research laboratory of Prof. Glass. The
solution reduction potentials have been determined by cyclic voltammetry, the electron
spin delocalization of the reduced complexes has been measured by EPR
spectroelectrochemistry, and the electronic structures have been modeled by DFT
computations.
54
Figure 3.1 Quinone complexes examined in this study.
55
Results and Discussion
Structure.
The synthesis of benzoquinone complex 7a has been reported previously.65
Another member of the research group employed both this procedure as well as a
different synthetic route, which can be found in more detail in the Organometallics paper
associated with this chapter.58 Consequently the synthesis will not be discussed here.
An ORTEP drawing of the molecular structure of 8b is shown in Figure 3.2. An
interesting feature about the structure of 8b is that the quinone moiety is face–to–face
with a phenyl ring from each of the triphenylphosphine ligands with 3.5884(10) Å
separating the centroid of each ring plane. Since such stacking may be due to π–π
interactions bis–phosphine complexes 8c and 8d were synthesized. The idea is that the
more electron–rich dialkoxyphenyl ring should preferentially stack. The structure of 8c
was determined by single crystal X–ray diffraction analysis and the ORTEP drawing is
shown in Figure 3.2. Indeed the more electron–rich phenyl rings preferentially stack with
the quinone moiety but the ring planes shift to a slipped interaction rather than the
eclipsed interaction seen in 8b. Rather than π–π interaction, electrostatic interaction
appears to be the preferential interaction of the methoxy–phenyl rings of the ligands with
the quinone e.g. the C18(C ortho to MeO and P of 9b ligand to C20(C=O) of
naphthoquinone) is 3.363 Å. Phosphine substitution reactions were carried out with
anthraquinone 10a. Reaction with phosphines 9b and 9c and Me3NO produced bis–
phosphines 10b and 10c.
56
Figure 3.2 Single crystal structure of 8b, and 8c with displacement ellipsoids at the 50%
probability level. The structure has two–fold rotational symmetry; the asymmetric unit is
the half of the molecule which is numbered.
57
Figure 3.3 Organic quinones in this study.
58
Electrochemistry.
The quinones attached to 2Fe2S cores are expected to be reduced at the more positive
potentials than their parent organic quinones due to increased delocalization/conjugation
and polarization effects. This is indeed the case as is shown in Figure 3.4, which displays
the voltammograms of the quinone–containing catalysts, 7b, 8a and 10a along with those
of the corresponding quinones 11b, 12 and 13. In metal complexes with non-innocent
ligands reduction can occur at the metal or ligand sites66,67 giving rise to species which
behave analogously to the separated species. Much attention has been devoted studying
non-innocent systems.33,68,69 The similarity between the behavior of the quinone–
containing metal complexes and the quinones is striking. Each voltammogram shows two
reversible one–electron processes corresponding to reduction of the quinone moiety to
the anion radical and dianion. This is unlike other enedithiolateFe2(CO)6 complexes such
as 4, which undergo a reversible two–electron reduction.15,70 The reduction pathway for
the quinone in the complexes is not changed by the presence of the S2Fe2(CO)6 portion of
the molecule. The sole effect is a substantial positive shift of the quinone potentials in the
complexes, +0.402–0.468 V for E1 and +0.223–0.416 V for E2 (results summarized in
Table 3.1). Thus, the S2Fe2(CO)6 portion acts as a very strong electron–withdrawing
substituent on the quinone. This substituent effect indicates that the S2Fe2(CO)6 “group”
is roughly equivalent to three chlorine atoms as judged by the observation that the
potential for one–electron reduction in acetonitrile of trichloro–1,4–benzoquinone is 0.43
V less negative than that of 1,4–benzoquinone itself.71 Similarly, the dianions produced
by electrochemical reduction of 7b, 8a and 10a render reduction of the S2Fe2(CO)6
portion of the molecule more difficult. Thus, the overall two–electron reduction potential
for the benzenedithiolato complex 4 is –1.47 V vs Fc+/Fc in DCM, but as Figure 3.5
59
shows further reduction of the quinonedithiolato complexes 7b2–, 8a2, 10a2– is not
observed in the potential range of these scans.
60
Figure 3.4 Voltammograms of ca. 0.5 mM quinone complexes 7b, 8a and 10a as well as
the corresponding quinones: 11b, 12 and 13. The currents have been divided by the
concentrations of the compounds to normalize for slight variations in concentration.
Dichloromethane with 0.10M n–Bu4NPF6 glassy carbon working electrode, 0.10 V/s.
61
Thus the quinone–containing complexes are easily reduced, even more so than the
free quinones, with E1 values falling in the range of –0.626 to –0.739 V vs. Fc+/Fc. It
seems possible that such complexes could catalyze the reduction of protons to form
hydrogen through a series of electron transfers and protonations at iron. However, even
with relatively weak acids such as acetic acid (pKa = 22.372 in acetonitrile), protonation
occurs at the quinone moiety, probably at the stage of the anion radical. Figure 3.5 shows
voltammograms of 0.49 mM 8a obtained with 0, 1 and 5 mM of added acetic acid. With
no acid, the voltammogram is the same as seen in Figure 3.4. Addition of 1 mM acid
causes a small increase in the first reduction peak and the second reduction peak is
replaced by a much larger quasireversible reduction peak near –1.5 V. As 1 mM acid is
stoichiometrically sufficient to convert the quinone–ligand to its hydroquinone form, the
new peak is assigned to reduction of the complex with a hydroquinone ligand, in
agreement with earlier results.70 Finally, in the presence of 5 mM acid, the first reduction
peak is still larger, the new peak for reduction of the hydroquinone–containing catalyst
becomes higher and sharper and a new peak appears near –1.9 V. As previously
reported,70 this peak corresponds to the reduction of acetic acid to dihydrogen and acetate
as catalyzed by the hydroquinone–containing complex. The catalytic peak is not present
until the ratio of acetic acid concentration to that of the complex exceeds two for it is
only then that there is excess acetic acid in the diffusion–kinetic layer. For ratios less than
or equal to two, the acetic acid is consumed to form the hydroquinone.73
62
Table 3.1 Standard potentials of quinone complexes in comparison to the corresponding
quinones.
Complex
11ba
7ba
11cc
2cc
11dc
7dc
E1 /V
–1.028
–0.626
–1.041
–0.552
–1.083
–0.661
E2 / V
–1.457
–1.234
–1.587
–1.202
–1.640
–1.315
ΔE1 b /V
–
–0.402
–
–0.489
–
–0.422
ΔE2 b /V
–
–0.223
–
–0.385
–
–0.325
–0.794
–1.469
–
–
11ec
c
7e
–0.424
–1.240
–0.370
–0.224
12a
–1.174
–1.675
–
–
a
8a
–0.718
–1.293
–0.456
–0.382
8bc
–1.082
–1.714
–0.092
0.039
a
13
–1.207
–1.667
–
–
10aa
–0.739
–1.251
–0.468
–0.416
a
Determined by simulations of cyclic voltammograms obtained in 0.10 M n–
Bu4NPF6/CH2Cl2 with a 0.3 cm diameter glassy carbon working electrode. Potentials
referred to the Fc+/Fc potential also determined in dichloromethane. The quinone–
containing complexes were not stable in acetonitrile. Diffusion coefficients used in
simulations are not reported as there was uncertainty in the concentrations due to
evaporation of solvent.
b
Potential of quinone minus potential of quinone complex.
c
Determined by averaging the potential of maximum current for the anodic and cathodic
peaks.
63
Data for multiple small increments in concentration from 0 to 15 mM show a smooth
transition from quinone–like electrochemistry to voltammograms showing formation of
the hydroquinone–containing complex and the subsequent catalytic reduction of the acid.
64
Figure 3.5 Voltammograms of 0.48 mM 8a with additions of 0, 1 and 5 mM acetic acid.
Mercury–film working electrode. Other conditions as in Figure 3.4.
65
As discussed previously, reduction of the quinone S2Fe2(CO)6complexes in the presence
of acetic acid results in formation of the corresponding hydroquinone complexes. Thus
the negatively charged oxygen in the semiquinone and/or quinone complex dianion is
protonated rather than the 2Fe2S moiety. However, this does not resolve a key hypothesis
of this approach; that is, delocalization from the quinone moiety to the 2Fe2S moiety on
reduction will occur (delocalization might occur but oxygen is preferentially protonated
leading to hydroquinone formation).
EPR Spectroscopy
To test whether delocalization occurs in the semiquinone, EPR spectroscopic
measurements were performed on the electrochemically produced semiquinone
complexes from 7b,c,d,e, 8a and 8b (Figure 5) and compared with the spectra obtained
from the parent semiquinones of 11b,c,d,e and 12. The spectra obtained for the
semiquinones of the parent quinones agree well with those reported in the literature.74
The gav–values and hyperfine splitting constants obtained for all of these semiquinones,
under the same conditions (aH values are known to change with solvent),74-76 are reported
in Table 3.2. Hyperfine and g-values for complicated splitting patterns were extracted
through simulation with the EasySpin software package49 and an example overlay of a
simulated and experimental spectrum can be seen in Figure 3.6. Parameters for the
EasySpin simulations can be found in Appendix B.
66
Figure 3.6 Simulated (green) spectrum overlaid with the experimental (black) spectrum
for 11b. The deviation from agreement can be seen in blue.
67
Figure 3.7 EPR Spectra for the semiquinone obtained on electrochemical reduction of
7b-e, 8a, 8b.
68
Comparison of the parent semiquinone g–values with that of the corresponding complex,
i.e semiquinone 11d vs 7d, 11e vs 7e and 12 vs 8a shows an increase in gav values for the
complexed semiquinone from 2.008 to 2.016–2.017. This suggests greater spin–orbit
coupling, attributable to delocalization of the unpaired electron from the quinone moiety
to the 2Fe2S moiety. In addition, comparison of the aH values for the uncomplexed
semiquinones with the corresponding 2Fe2S complexed species shows approximately a
40% decrease in aH for the complexed moieties. This provides further support for
delocalization of the unpaired electron from the quinone to the 2Fe2S moiety.
Comparison of the EPR spectra of the semiquinone obtained from 8a with that from the
bis-triphenylphosphine substituted, 8b, shows a dramatic difference. The former shows a
broad singlet but the latter a distinct triplet. The additional hyperfine is clearly due to
splitting by the two equivalent
31
P atoms of the Ph3P ligands, irrefutably demonstrating
spin density on the P2Fe2S2 moiety. The observed
31
P hyperfine is 4.1 MHz, which is
much smaller than 31P splitting found in other phosphine substituted hydrogenase mimics.
77-79
DFT calculations were performed on 8b– in order to investigate the cause of the
smaller hyperfine value than literature values for similar cationic compounds.77-79
Examination of the SOMO of 8b– semiquinone shown in Figure 3.8 identifies the reason
for the modest hyperfine splitting by 31P. That is, the metal orbital that is participating in
the SOMO has a node at the apical position occupied by P. This prevents delocalization
of the SOMO onto the P ligands and thus the phosphines split solely due to spin
polarization and not due to spin delocalization via orbital mixing. Nevertheless,
delocalization of the unpaired electron from the quinone to 2Fe2S moiety is established.
Computations
69
DFT studies were also carried out on the other semiquinones and semiquinone
complexes studied and the empirical hyperfine constants compare favorably. Agreement
of the calculated aH, shown in Table 3.3, with the experimental values shown in Table 3.2
validates the computations. DFT calculations were performed using both ORCA51 and
ADF. The two programs were compared extensively and found to be near equivalent.
ADF was chosen as the software package due to the use of Slater-type orbitals. As shown
in Table 3.3 the calculated spin density on the carbon α to the remaining proton in
7b,c,d,e shows reduction of spin density in the complexed semiquinones as compared to
the uncomplexed semiquinones. Furthermore the calculations show substantial spin
density on Fe in the complexed semiquinones, thereby providing theoretical support for
extensive delocalization in the complexed semiquinones. Additionally the amount of spin
density on the α C and Fe centers (16% to 23% for Fe) can be tuned by varying the
functionalization of the quinone moiety.
70
Table 3.2 EPR Spectroscopic Parameters of Complexed and Uncomplexed
Semiquinones.
Semiquinone
gav
aH/Pa
Precursor
(MHz)
2.009
–2.5
11b
2.0177
–1.1
7b
2.008
–5.32
11c
2.017
–2.80
7c
2.008
–5.5
11d
2.017
–3.1
7d
2.008
–6.9
11e
2.016
–4.4
7e
2.008
–1.8
12
2.017
–
8a
2.017
–4.1
8b
a
All a values reported are for proton connected to C5 of the quinone ring
except for compound 3b whose reported a value is for the phosphorus
hyperfine.
71
Conclusions
The electrochemical behavior of the quinone complexes is similar to that of the free
quinones, but with a considerable positive shift in the reduction potentials due to both
increased overlap and polarization effects as indicated by DFT calculations. The quinone
dithiolate ligand was demonstrated to be non–innocent through cyclic voltammetry, EPR
spectroscopy, and DFT computations. Changing functionalization of the quinone moiety
enabled tuning of the spin density on the metal centers; Fe spin density could be changed
from 16% to 23 % by changing from chloroquinone to naphthoquinone, with
intermediate values for methyl, methoxy, and tert–butyl 1,4–benzoquinone Fe2S2(CO)6
complexes. Upon reduction in the presence of acid the hydroquinone product is formed
which can then in turn catalytically produce molecular hydrogen.
70
These systems
demonstrate that unsaturated π–systems behave non-innocently in their electronic
properties, and functionalization of the unsaturated π–system can influence electronic
structure.
72
Figure 3.8 Calculated SOMO of 8b—.
73
Table 3.3 Calculated EPR Spectroscopic Hyperfine Splitting and Mulliken Spin
Densities.
Semiquinone
aH
Mulliken Spin Density %
Precursor
(MHz)
Fea
Cb
5.1
–
4.7
11b
3.5
18.32
1.4
7b
6.1
–
6.2
11c
3.3
18.8
2.8
7c
5.5
–
6.3
11d
3.3
18.7
3.0
7d
6.7
–
8.9
11e
4.3
15.6
4.3
7e
1.7
–
2.0
12
–
22.7
0.8
8a
–
17.4
1.3
8b
a
The sum of both iron centers.
b
Mulliken spin density for the carbon atom in the 5 position of the parent quinone, and
the corresponding carbon in the naphthoquinone analogs.
74
Chapter 4 Comparison of the Electronic Structure of 1,2-(µbenzenedithiolato)-2’-phenylazopyridinediiron-tetracarbonyl and
1,3-(µ-propanedithiolato)-2’-phenylazopyridinediirontetraacarbonyl to 1,2-(µ-benzenedithiolato)-hexacarbonyl and 1,3(µ-propanedithiolato)-hexacarbonyl
Part of this chapter has been published in the Journal of Sulfur Chemistry.80
Introduction
A mimic of the hydrogenase active-site, which appended to a non-innocent
chromophore, could prove desirable in its potential ability to use energy captured from
light to assist in production of molecular hydrogen. As seen in Figure 1.3, the idea is that
once excitation from an occupied orbital to the LUMO of the chromophore takes place,
an electron can be inserted into the newly formed hole at a potential less negative than
needed for inserting directly into the LUMO of the molecule. 2-phenylazopyridine, 14,
has a bright orange color due to a π to π* transition, and is capable of ligating metal
centers. A Ru(II) hydride complex has been synthesized with 14 as a ligand which has
been shown to generate molecular hydrogen.81 Compound 14 is also known to be reduced
to its radical anion and dianion, opening the possibility of multiple redox states when
ligated to the metal which could facilitate the conversion of the metal from one redox
state to another. To this end, two well-known hydrogenase active-site mimics, 4
17-19
and 3a,
15,16,82
have been substituted with 14 in place of two carbonyl ligands to form 1,2-
(µ-benzenedithiolato)-2’-phenylazopyridinediiron-tetracarbonyl
(15)
and
1,3-(µ-
propanedithiolato)-2’-phenylazopyridinediiron-tetraacarbonyl (16) as seen in Figure 4.1.
Complexes 15 and 16 display an intense blue color, while their unsubstituted
counterparts, 4 and 3a are red. This drastic change in color suggests a charge transfer
event between the metal and ligand. To understand the electronic structure of these
75
complexes, including the blue color, DFT studies were performed in conjunction with
spectroscopic measurements. DFT studies were validated by comparison of the calculated
data with that empirically found from single crystal X-ray diffraction, IR spectroscopy,
and UV-Vis spectroscopy.
This chapter will show that molecular orbitals of 15 and 16 change considerably
upon complexation of 14 and have extensive delocalization between the diiron core and
the 2-phenylazopyridine ligand, as well as exhibiting a charge transfer event from the
metal centers to the LUMO of the 2-phenylazopyridine ligand. The redox state of 14
when ligated to 15 and 16 is also investigated. The molecules studied here were
originally synthesized by other members of the CRC group and consequently the
synthesis will not be discussed herein. The synthetic route has been presented in detail.80
Additionally it should be noted that some of the UV-Vis spectral data presented herein
was collected by Matthew Swenson.
76
Figure 4.1 The four primary molecules of interest in this chapter.
77
Comparison of structural parameters:
Single crystal X-ray diffraction was performed on both complexes by Dr. Gary S.
Nichol. The crystal structures of 15 and 16 have the 2-phenylazopyridine ligand bound in
a basal-basal conformation to a single iron. For 16 the central carbon of the dithiolate
bridge pointed away from the 2-phenylazopyridine ligand. DFT computations were
performed on multiple conformations of 16 to search for other isomers which may be
more stable in the gas-phase and possibly in solution. Conformations were tested with the
2-phenylazopyridine ligand binding in multiple fashions to one iron center as well as
bridging the two iron atoms. It was found that the lowest energy conformation matched
the single-crystal X-ray structure. Flipping of the bridge resulted in a structure which was
1.6 kcal/mol higher in energy in solution with an equilibrium constant of 6.6x10-2. All
other isomers were substantially higher in energy. A comparison of key bond lengths
between the calculated and x-ray structures can be found in Table 4.1. All bond lengths
were well reproduced, within 0.01 Å in most cases. The slight increase in N=N bond
length of ~0.05 Ǻ from 1.25 ± 0.02 Å typical of uncoordinated molecules of this type61 to
1.311 (3) Å in 16 is also reproduced by the computations, and shows that the 2phenylazopyridine ligand retains double bond character suggesting a neutral ligand.
IR spectroscopy: CO stretching frequencies
Experimental and calculated CO stretching frequencies were compared to each
other (Figure 4.2). The calculations match the experimental CO region of the IR spectrum
well, suggesting the structural functional and basis set are adequately handling the
electronic structure of 15 and 16 appropriately. The calculated stretching frequencies of
15
and
16
were shifted by a factor
of 1.003
and
1.005,
respectively.
78
Figure 4.1 Comparison of the experimental IR spectrum (black) and calculated spectrum
(blue) in the metal-carbonyl stretching frequency region for 15 (above) and 16 (below).
79
Table 4.1 Comparison of key bond lengths for 15 and 16 between the X-ray crystal
structure and DFT calculations.
15
Fe1-Fe2
Fe1-S
Fe2-S
Fe1-N1
Fe1-N3
N1-N2
16
Fe1-Fe2
Fe1-S
Fe2-S
Fe1-N1
Fe1-N3
N1-N2
X-ray Distance (Å)
Calculated Distance (Å)
2.49
2.26
2.27
1.89
1.93
1.31
2.49
2.26
2.28
1.87
1.91
1.31
X-ray Distance (Å)
Calculated Distance (Å)
2.52
2.23
2.27
1.88
1.94
1.32
2.53
2.23
2.27
1.89
1.93
1.31
80
UV-Vis spectroscopy
Compounds 15 and 16 have an intense blue color, while 4, 3a, and 14 are
orange/red, leading to a drastic change in the electronic spectra of 15 and 16 by
comparison to their parent hexacarbonyl analogs. This difference can be seen in Figure
4.3 with the appearance of an intense absorption at nominally 600 nm with an ϵ for the
λmax approximately 8 times larger than the λmax of the parent hexacarbonyl compounds.
The UV-Vis spectra were collected in a variety of solvents of different polarities to test
the possibility of λmax dependence upon solvent polarity. The dependence on solvent
polarity would arise due to the change in dielectric constant causing a change in the
difference between the solvation energies of the ground and excited states. Minimal
change in λmax was observed in a wide variety of solvent polarities. The absorption bands
for 15 and 16 are: 600 nm, 603 nm, 600 nm, 600 nm, and 604nm for 15, and 605 nm,
608 nm, 602 nm, 605 nm, and 609 nm for 16) in THF ( = 9) CH2Cl2 ( = 8), EtOH ( =
25), CH3CN ( = 36), DMSO ( = 37), respectively. No correlation can be made in the
shift of solvent polarity to λmax suggesting no charge transfer event was present; the
magnitude of the dipole moment is not significantly changed upon charge transfer; or
solvent molecules were not able to access the portions of the molecule affected by the
dipole moment change.
81
Extinction Coefficient(cm-1M-1)
15000
10000
A
5000
0
375
425
475
525
575
625
675
725
775
Wavelength (nm)
Extinction Coefficient(cm-1M-1)
15000
10000
B
5000
0
375
425
475
525
575
625
675
725
775
Wavelength (nm)
Figure 4.2 A.UV-Vis spectrum of 15 (blue) 4 (purple) and 14 (red) in dichloromethane.
B.UV-Vis spectrum of 16 (blue) 2a (purple) and 14 (red) in dichloromethane.
82
DFT computations and correlation diagrams
DFT computations, which were validated by IR spectroscopy and single crystal
X-ray structure, were performed to gain further insight into this absorption and the
electronic structure of this complex. The calculated HOMO, LUMO, and LUMO +1 of 4
and 15 are shown in Figure 4.4. The HOMO is almost entirely localized on the
substituted Fe with contribution from this Fe totalling 75% of the orbital character, and
only 9 % orbital character on the other Fe. Only 4 % of the orbital character is distributed
between the two S atoms. This is in contrast to the higher symmetry molecule, 4, in
which the HOMO contains considerable Fe-S mixing with 39 % distributed equally
between the two Fe atoms, and 18 % distributed equally among the two S atoms, and the
remainder ascribed to aromatic π character. This preferential donation of orbital character
from one half of the molecule, and minimal 2-phenylazopyridine π character, is seen in
all occupied orbitals from the HOMO through HOMO-6, which correspond primarily to
the three occupied d-orbitals on each metal and the metal-metal bond. The HOMO energy
of 15 is raised relative to 4 due to the loss of π back-bonding to the two carbonyl ligands
that 2-phenylazopyridine replaces (Figure 4.5) and removal of delocalization to the S
atoms. The 2-phenylazopyridine ligand has some slight π-backbonding ability such that
the σ-donation of charge from the ligand is compensated and the overall ligand is near
neutral. The 2-phenylazopyridine acceptor orbital has a π-antibonding symmetry between
the azo-N atoms and the weak π back-donation from the metal, consistent with the
observed ~0.05 Ǻ increase in the N=N distance from the free ligand molecule mentioned
earlier.
83
Table 4.2 Calculated solvated energies (eV) in CH2Cl2, CH3CN and CH3CH2OH, of
various singlet excited states to the LUMO of compound 15.
Hole Location
CH3CN
-314.60
-313.12
-312.97
-312.28
-312.12
NA/ Ground State
HOMO
HOMO -1
HOMO -2
HOMO -4
Δmaxa
Solvent
DCM
-314.54
-313.06
-312.91
-312.21
-312.05
EtOH
-314.59
-313.114
-312.96
-312.63
-312.11
0.00
0.00
0.37b
0.01
a
This was calculated by taking the largest difference between the ground state and excited
state for a given excitation and subtracting it from the smallest difference for a given
excitation.
b
Significant rehybridization of the excited state orbital character in EtOH has caused this
point to be an outlier.
Table 4.3 Calculated solvated energies (eV) in CH2Cl2, CH3CN and CH3CH2OH, of
various singlet excited states to the LUMO of compound 16.
Hole Location
NA/ Ground State
HOMO
HOMO -1
HOMO -2
HOMO -3
HOMO -4
a
Δmaxa
Solvent
CH3CN
-295.25
-293.87
-293.66
-293.45
-292.97
-292.86
DCM
-295.20
-293.80
-293.60
-293.39
-292.90
-292.79
EtOH
-295.24
-293.85
-293.64
-293.44
-292.96
-292.85
0.01
0.01
0.01
0.01
0.01
This was calculated by taking the largest difference between the ground state and excited
state for a given excitation and subtracting it from the smallest difference for a given
excitation.
84
Figure 4.3 LUMO +1 (top) LUMO (middle) and HOMO (below) calculated for 4 (left)
and 15 (right).
85
Figure 4.4 Orbital energy correlation diagram. (a) Orbital energies of parent
hexacarbonyl compound (the “2Fe2S” is a block of seven orbitals that derive from the
two d6 Fe centers and the Fe-Fe bond). (b) The “2Fe2S” fragment orbital energies. (c)
Orbital energies of overall compound with the numbers leading to the LUMO showing
the percent contribution of contributing fragment orbitals (d) LUMO of 14.
86
The LUMO of 4, similar to the HOMO, has a distribution of orbital character
between the Fe atoms and the S atoms with 44 % Fe character and 18 % S character.
Incorporation of the 2-phenylazopyridine ligand raises the energy of the LUMO of 4 so
that it now forms the LUMO+1 of 15, retaining a strong resemblance to the LUMO of 4
with the addition of 2-phenylazopyridine character. The new low-lying empty orbital
introduced by 2-phenylazopyridine mixes with the LUMO +1 of 4 to form the LUMO of
molecule 15 (66% character from 2-phenylazopyridine). The antibonding interaction
between the azo-N atoms alone accounts for 30% of the LUMO of 15. The LUMO of 15
includes some mixing and character traced to the HOMO and LUMO+1 of 4, resulting in
the Fe atoms contributing 16 %, and 9% and almost eliminating the contribution of the S
atoms (3 %) in the LUMO. This mixing and increased electron richness in the 2Fe2S core
when two carbonyls are substituted by 2-phenylazopyridine results in the LUMO’s of 4
and 15 having nearly the same energy.
Gas-phase time-dependent DFT (TD-DFT) computations predict a strong
absorption at 578 nm for 15, in good agreement with the observed transition at 598±3
nm. The shift of this absorption to longer wavelength from the first significantly allowed
transition of 4 is calculated to be 135 nm in good agreement with the observed shift of
145±5 nm. The TD-DFT computations describe this absorption of 15 as excitation into
the LUMO from 0.64 HOMO-2, 0.14 HOMO-1, and 0.10 HOMO-4. The longer
wavelength of the first absorption for 15 compared with 4 is ascribed to a closing of the
gap between the LUMO and the 2Fe2S occupied orbitals in 15 due to raising of the
2Fe2S occupied orbital energies (Figure 4.5) compared with 4. Because the highest
occupied orbitals are primarily metal based and the LUMO is primarily 2-
87
phenylazopyridine based, this absorption is defined as a metal-to-ligand charge transfer
transition. The calculated oscillator strength for excitation to the LUMO is five times
greater for 15 than for 4, which is reasonable in comparison to the eightfold increased
extinction coefficient for this transition observed in the experimental spectrum. The
assignment of the absorption to a charge-transfer band is consistent with the large
extinction coefficient as observed, but in apparent contradiction to the lack of λmax shift
with solvent polarity. However, the large size of these molecules and the fluid electron
density provided by the 2Fe2S core and carbonyl ligands may minimize the ability of the
solvent to differentially stabilize the ground and excited electronic states. As a test of this
possibility, the difference in solvation energy stabilization between the ground state and
various contributing excited states were calculated via DFT for a variety of solvents with
a range of dielectric constants; CH2Cl2 ( = 9), EtOH ( = 25), and CH3CN ( = 36). The
solvation energy differences were found to produce shifts of no more than 3 nm, which is
within the uncertainty of the experiment, and a summary of the solvation energy changes
can be seen in Table 2. Thus the computations show a strong charge-transfer absorption
at long wavelength that does not shift significantly with solvent polarity, in agreement
with the experiment. Apparently the absorption wavelength dependence on polarity is a
poor test for a charge transfer transition for this class of molecules.
Compounds 3a and 16 show much of the same trends as 4 and 15. The N-N
distance in 16 complex is 1.315(2) Å which is again consistent with a N=N bond slightly
weakened by backbonding from the metals and further supported by the N=N stretching
frequency seen in the IR spectrum. As with 4 and 15, the UV-Vis absorption spectrum for
16 showed a new band at long wavelength (605 nm), shifted from that seen in 3a.
88
Figure 4.5 LUMO +1 (top) LUMO (middle) and HOMO (below) calculated for 3a and
16.
89
This absorption band shifted only modestly (605 nm, 608 nm, 602 nm, 605 nm,
609 nm) in THF, CH2Cl2, EtOH, CH3CN, DMSO respectively and as with 15 did not
correlate with solvent polarity, and the results of DFT calculated solvation energies for
select solvents are summarized in Table 4.3. DFT calculations again suggest the shift in
absorption for 16 compared to 3a is attributable primarily to a raising of the HOMO as
can be seen in Figure 4.5. The composition of the HOMO of 16 is very similar to that of
15 with 75% of the orbital character coming from the Fe ligated with 2phenylazopyridine. This is in contrast to the HOMO comparison of 3a to 4, in which the
HOMO character is very much different between the two complexes. The high energy
occupied orbitals remain primarily 2Fe2S in character, and the 2-phenylazopyridine
ligand LUMO comprises 64% of the LUMO of 16. According to TD-DFT calculations
the long wavelength absorption in 16 is from the HOMO-2 (0.36) HOMO-1 (0.26) and
HOMO-3(0.22) to the LUMO. The key orbital compositions are illustrated in Figure 4.6.
90
Conclusions
The electronic structure of two new diiron complexes of 2-phenylazopyridine have been
studied. Introduction of 2-phenylazopyridine as a chelate ligand on one Fe in place of two
CO ligands results in charge asymmetry in the 2Fe2S core and localization of high energy
occupied orbitals on the Fe centers. In addition the LUMO of the substituted complexes
shows extensive contribution from the 2-phenylazopyridine ligand π orbital. A
consequence of these changes is that an absorption band is observed at long wavelength
(ca. 600 nm) with a substantial increase in extinction coefficient in the visible absorption
spectrum of these complexes. Time-dependent DFT calculations show that an absorption
results from excitation from the filled Fe orbitals to the low-lying empty 2phenylazopyridine orbital. Polar solvents are not able to provide significant additional
stabilization of the charge-transfer excited state than is provided within the electronic
structure of the molecule itself. The overall picture that emerges is one of fluid orbital
distributions in the 2Fe2S core aided by the nature of sulfur orbital interactions with iron
atoms and the backbonding capabilities of carbonyls. The mixed character and energy of
the LUMO in the 2-phenylazopyridine complexes indicates that this ligand will be noninnocent in the reduction chemistry of these complexes, and the deep blue color suggests
the possibility of interesting excited state chemistry.
91
Chapter 5 Persistence of Disulfide Bonds in Reduced Bipyridines
and the Observance of Intramolecular Charge Transfer.
Part of this chapter has been published in the Journal of the American Chemical
Society.83
Introduction
Electrochemical reduction of compounds with two reducible moieties presents some
interesting issues. Is each moiety reduced independently of the other or is there
interaction? This interaction may result from orbital overlap and delocalization of the
added electron or by internal electron transfer between orbitals centered on each
reducible group.84,85 This paper examines the electrochemical behavior of disulfides 17,
18+ and 19a2+ and studies of their electrochemical reduction. Compounds 17, 18+, and
19a2+ are 1,2−dithiins and the electrochemistry of a number of other 1,2−dithiins has
been investigated86-91 but those studies have emphasized oxidations, while the current
study will focus on the reduction of these compounds. The reduction of 17, 18+ and 19a2+
are of particular interest because they can be viewed as having two sites for reduction: the
bipyridine/pyridinium and disulfide moieties which may interact with each other.
92
Scheme 5.1 Relevant compounds to this chapter.
93
The electrochemical reduction of aromatic disulfides has been much studied including
investigations of the nature of the cleavage of the S−S bond on the path to eventual
formation of two thiolates, reactions 1–3.92-95 The cleavage is normally thought to occur
at the level of the anion radical as a very fast reaction. However, the anion radical does
have a finite lifetime, unlike the reduction of peroxides which proceeds by a dissociative
electron−transfer reaction giving in a single step the alkoxyl radical and the alkoxide.96
Ar−S−S−Ar + e− ⇌ [Ar−S−S−Ar] • −
(1)
[Ar−S−S−Ar]• −  Ar−S− + Ar−S•
(2)
Ar−S• + e− ⇌ Ar−S−
(3)
The reason for the short lifetime of the disulfide anion radical is that there is a 2c,3e (σ2
σ*1) bond between the sulfur atoms in which the σ*–electron weakens the S–S bond
resulting in its elongation.97 In fact, the inner reorganization energies are consistent with
considerable lengthening of that bond in the anion radical compared to the neutral
disulfide.92
Disulfide anion radicals have attracted much interest because of their biological
relevance. It has been suggested98,99 that in Class I ribonucleotide reductases a disulfide
anion radical reduces a 3′−ketodeoxynucleotide provided that a nearby glutamate residue
(E441) is protonated. Ribonucleotide reductases catalyze the reduction of the 2′–OH
group of ribonucleotides to 2′–deoxyribonucleotides required for DNA synthesis. In
addition, Erv2p, a sulfhydryl oxidase which catalyzes the synthesis of protein disulfides
from dithiols and oxygen, may shuttle electrons from protein dithiols via a disulfide anion
94
radical from a disulfide
nearby
additional
on a flexible arm to the FAD redox site with a
disulfide bond.100 Since disulfides are important
in maintaining protein and peptide structures, cleaving such bonds by addition of
electrons has also been a focus of attention. Both chemical and theoretical studies on
protein, peptide and model compounds have been performed to determine the stability of
the disulfide radical ions and the relative susceptibility of disulfides to adding an electron.
Of particular relevance to the studies reported here, it was found that incorporating
disulfides into rings which results in ring strain, favors formation of the disulfide anion
radical101 as does nearby positive charge.102,103 Topological constraints in peptides,103
proximity of amide groups and protonation in proteins were found to stabilize disulfide
anion radicals.104 Cleavage of the S−S bond in disulfide anion radicals in which the
disulfide is incorporated in ring systems is less favorable due to entropic effects than in
acyclic systems in which the resulting thiolate and thiol radical can diffuse apart. In a
series of papers on the formation and cleavage of aromatic disulfide anion radicals the
effect of substituents was investigated.92-96 In addition, it was shown that the reaction
involves a stepwise dissociative electron−transfer with significant inner reorganization as
pointed out above.
There are even constrained examples in which the anion radical of a disulfide either
retains the S−S bond or the cleavage reaction is reversible. The anion radical of
substituted 1,2−dithiin 20, which has the same 1,2−dithiin moiety as 17, 18+ and 19a2+,
may have an intact S−S bond though isolation of solid (Ph4P)(C6S8) showed that it was
actually a dimer, (Ph4P)2(C12S16), with the original S−S bond broken and a new S−S bond
being present between the original monomeric components of the dimer. In solution,
95
C12S162− is in equilibrium with the monomer, C6S8•−, with an equilibrium constant that
indicates that dilute solutions contain large fractions of the anion radical. Cyclic
voltammetry of such solutions shows two reversible redox processes which have been
interpreted as being due to the C6S8•−/C6S82− and C6S8/C6S8•− couples. It is not clear
whether the S−S bond persists in the anion radical, but it may.105 In addition one−electron
reduction of 21 gives the corresponding anion radical which is resistant to further
reduction in contrast to most diaryl disulfides.106 Furthermore this disulfide anion radical
or thiolate−thiyl is stable under cyclic voltammetric conditions. The disulfide anion
radical obtained from uracil disulfide, 22, by addition of hydrated electrons in a radiation
chemical study was more stable toward dissociation than other diaryl disulfide anion
radicals.107 This stabilization toward dissociation was ascribed to favorable stacking of
the heterocyclic rings.
The electrochemistry of N,N′−dialkyl−4,4′−bipyridinium salts, viologens, has been
extensively investigated because of their importance as herbicides,108 electron relays,109
redox mediators110,111 and use in molecular devices.112,113 Typically they show two low
potential reversible one−electron reductions. These potentials depend significantly on the
substituents in the bipyridinium rings but not the N−alkyl groups.114 In geometrically
restricted viologens in which there is a bridge of varying length between the 3 and
3′−positions, the dihedral angle between the two planar pyridinium rings roughly
correlates with the first reduction potential.114
Recently, it has been reported115 that a close analogue of 19a, that is, 19b, undergoes
four reversible one-electron reductions in which two electrons are added to the
bipyridinium moiety followed by two-electron reduction of the disulfide moiety to give
96
the ring opened dithiolate. In our studies of the reduction of 19a by cyclic voltammetry
there are four reversible one-electron reductions analogous to the electrochemical results
with 19b, but our calculations show that the mechanism for these reductions is different
from that suggested for 19b. Specifically, we find a fascinating interplay between the
redox moieties featuring internal electron transfer. In addition, the reduction of the
disulfide moiety occurs with potential inversion resulting in a two-electron reduction
concomitant with disulfide cleavage.
Results and Discussion
Behavior of 17
Cyclic voltammograms of 17 in acetonitrile with a glassy carbon electrode at various
scan rates is shown in Figure 5.1. As ascertained by the decreasing normalized current
with increasing scan rate, there is an overall two−electron reduction on the initial
negative−going scan at -1.34 V vs. Fc+/Fc and an electrochemically separated
two−electron oxidation on the return scan. Scans to −2.7 V did not reveal any additional
reduction processes. The overall two−electron stoichiometry is supported by the match of
the simulation to the experimental curve (shown in Figure 5.3), as well as DFT
computations that calculate potential inversion with the reductions occurring at -1.34 V
and -1.02 V giving an Eov of -1.18 V, close the simulated value of -1.13 V. While the
electrochemical simulations were performed over a wide range of scan rates, it should be
emphasized that the parameters used to simulate the cyclic voltammogram are probably
not unique in their ability to produce a fit. This can be seen when examining
Figure 5.3 and Figure 5.2. They have been simulated by two different mechanisms
(Scheme
5.2
and
Scheme
5.3),
and
both
give
reasonable
fits.
97
-200
4, 4'-bipyridine 3, 3'-disulfide (1) acetonitrile
0.97mM, GCE
50 mV/s
100 mV/s
-150
200 mV/s
300 mV/s
In-0.5 / A/(V/s)0.5
-100
500 mV/s
1 V/s
-50
2 V/s
3 V/s
5 V/s
0
10 V/s
20 V/s
50
30 V/s
100
0.00
-0.50
-1.00
-1.50
E / V Vs.
-2.00
-2.50
Fc+/Fc
Figure 5.1 Cyclic voltammograms of 0.97 mM 17 in 0.10 M Bu4NPF6/acetonitrile at 295
K. Glassy carbon working electrode.
98
0.10 V / s
50
I/A
30
10
-10
-30
0.0
-0.5
200
-1.0
E / V Vs. Fc+ / Fc
-1.5
-2.0
-2.5
-1.5
-2.0
-2.5
-2.0
-2.5
1.0 V / s
150
I/A
100
50
0
-50
-100
0.0
-0.5
-1.0
E / V Vs. Fc+ / Fc
600
10 V / s
500
400
I/A
300
200
100
0
-100
-200
-300
0.0
-0.5
-1.0
E / V Vs. Fc+ / Fc
-1.5
Figure 5.2 Experimental cyclic voltammogram of 17 under the same conditions as
present in Figure 5.1, and simulation (black circles) according to Scheme 5.2. E4 = -1.35
V α4= 0.75, ks4= 0.003 cm/s, E5 = −0.90 V; 5 = 0.66; Dall species = 2.3  10−5 cm2/s; rdisk
= 0.15 cm. Subscript numbers refer to the chemical equations in Scheme 5.2. The
simulation is based on diffusion to a disk electrode.
99
250
0.10 V / s
200
I/A
150
100
50
0
-50
-100
0.0
-0.5
-1.0
-1.5
-2.0
-2.5
-2.0
-2.5
-2.0
-2.5
E / V Vs. Fc+ / Fc
250
1.0 V / s
200
150
I/A
100
50
0
-50
-100
0.0
-0.5
-1.0
-1.5
E / V Vs. Fc+ / Fc
700
10 V / s
600
500
400
I/A
300
200
100
0
-100
-200
-300
0.0
-0.5
-1.0
-1.5
E / V Vs. Fc+ / Fc
Figure 5.3 Experimental cyclic voltammogram of 17 under the same conditions as
present in Figure 5.1, and simulation (black circles) according to E6 = -1.435 V α6=
0.896, ks6= 0.052 cm/s, E8 = -0.891 V α8= 0.587, ks8= 0.003 cm/s, Keq7= 5.93, kf7=
11650, Dall species = 2.3  10−5 cm2/s; rdisk = 0.15 cm. Subscript numbers refer to the
chemical equations in Scheme 5.2. The simulation is based on diffusion to a disk
electrode.
100
Electrochemical reduction of the carbocyclic analog of 17 (nitrogen atoms
replaced by CH) has not been studied but the somewhat related diphenyl disulfide (PhSSPh) shows an irreversible reduction peak potential in DMF of − 2.11 V vs. Fc+/Fc.93
This is much more negative than the − 1.18 V vs. Fc+/Fc in acetonitrile for the standard
potential for 17 + 2 e− ⇌ 172− found in this work. Molecular mechanics calculations
(MM2) showed a difference of strain energy between the carbon analogue of 17 and its
corresponding ring−opened dithiol of −1.1 kJ/mol suggesting that relief of steric strain is
not important in these redox processes.116
Further insight into the addition of the first electron to 17 is obtained from the
computational results. The LUMO calculated for 17 is shown in Figure 5.5. Interestingly
it shows mixing of sulfur and π−orbitals of the heterocyclic rings. In addition, it shows
strong anti−bonding character between the sulfur atoms which would promote S-S bond
cleavage and rotation of the rings about the C4-C4’ bond; with a much smaller
contribution to the LUMO from a π bonding interaction between the 4 and 4' carbons that
discourages rotation. Thus the first electron upon reduction goes into an orbital
dominated by S−S antibond character, causing 17− to open the torsion angle by
approximately 10° from 30° to 40° for geometries optimized with acetonitrile
solvation (Figure 5.4). This lengthens the S−S bond and rehybridizes the SOMO to
almost exclusively S−S antibonding character and virtually no C4-C4’ π bonding (Figure
5.5).
101
Scheme 5.2 Proposed electrochemical mechanism for compound 17.
102
Scheme 5.3 "Conventional" mechanism of S-S bond cleavage.
103
A second shallow well for 17− is present 0.51 eV higher in energy at a torsion angle of
~100° with the sulfur atoms rotated away from each other. Upon reduction to 172- the
S−S bond breaks, and the sulfur atoms rotate away from each other to a torsion angle of
~100° with no calculated barrier to rotation. Thus, the disulfide anion radical of 17 is an
intermediate in this dissociative electron transfer as is the case with 20•− and 21•−.
However, due to potential inversion the anion radical of 17 rapidly reduces to the dianion.
Nevertheless, simulations of the experimental voltammograms over a wide range of scan
rates produced reasonable fits for both a 2-electron reduction concerted with S-S bond
cleavage and rotation, as well as a two 1-electron sequence followed by cleavage of the
S-S bond. Consequently, it is not possible to conclude from the experimental data the
degree to which the second reduction is coupled with geometric rearrangement; however,
the simulations of the former case produce the most reasonable electron transfer rates.
The lack of electrochemical reversibility but chemical reversibility, that is, 17 after
reduction can be reoxidized, can be explained by following the rotation energy profiles in
the reverse direction. Oxidation of 172− is slowed due to the aforementioned potential
inversion making it an irreversible process on the electrochemical time scale. This is
displayed in reactions 4−5, Scheme 5.2 with the corresponding simulation parameter
values given in the caption of Figure 5.3.
104
Neutral
-132
-134
-136
Energy (eV)
Anion
-138
-140
Dianion
-142
-144
0
50
100
S-C4-C4'-S' dihedral angle
150
Figure 5.4 Energy rotation profile, with respect to the S-C4-C4’-S’ torsion angle for 17
(top trace), 17- (middle trace) and 172- (bottom trace).
105
Figure 5.5 LUMO of 17 (top), SOMO of 17- (center) and HOMO of 172- (bottom).
106
Behavior of 19a2+
The positive charges in the rings of 192+ result in a calculated LUMO, shown in Figure
5.8, consisting mostly of heterocyclic ring π−character with minimal S−S antibonding
character compared with 17 (Figure 5.5). This suggests that upon reduction the first
electron will go into the π−system and not the S–S moiety as was the case with 17. This
is comparable to that suggested for the reduction of bis−p−nitrophenyl disulfide,92,94,95
where two energy minima were computationally identified: one with the unpaired
electron localized on a nitro group with only a slightly elongated S–S bond; and the other
localized on the S–S bond with an elongated S–S bond similar to other disulfide anion
radicals. Thus, cleavage of the S–S bond in these methylated bipyridinyl systems, may
involve internal electron transfer from the π−system to the S–S bond.
The cyclic voltammogram obtained for 19a2+ is shown in Figure 5.6. Two pairs of
voltammetric peaks are seen with the first pair occurring between -0.4 to -0.6 V. Note
that these reductions occur at potentials less negative than those found for 17. In addition,
these processes are fully reversible on the voltammetric time scale as attested to by the
good fit of the simulation to the experimental voltammogram. One also notes the second
pair of reversible reduction processes in the region of −1.7 to −2.0 V. Interestingly, the
reduction potentials of bipyridinium salts depend on the dihedral angle between the two
heterocyclic rings. In the solid state, this angle is 32° for 17,115 which closely matches our
calculated value of 34°. This is also close to the calculated value for 192+ of 36°. and is
considerably less than that for gas-phase calculated methyl viologen2+(44°).117,118 On
reduction, the dihedral angle is calculated to be close to 0° for the methyl viologen cation,
promoting electron delocalization between the two rings. Consequently reduction of
19a2+, should be more facile than methyl viologen2+ and there is also a substituent effect
107
on the reduction potentials for 19a2+. Indeed the first two reduction potentials for 19a2+
are less negative than the corresponding standard potentials for methyl viologen2+ of
−0.80 and −1.25 V.119 The standard potentials for the four steps of reduction of 19a2+ are
−0.49, −0.60, −1.79 and −2.05 V, which compare favorably with the DFT calculated
values of -0.55, -0.64, -1.75, and -2.16 V.
DFT calculations show the SOMO of 19a+ is primarily comprised of π−character with
the S-S bond intact, and having almost identical orbital contributions as the LUMO of
19a2+, i.e. almost no orbital rehybridization takes place upon the first reduction as seen
from Figure 5.8. With the SOMO of 19a+ being the site of the next reduction, initial
insertion of an electron into 19a is to the π−system. However, after the electron has been
accepted to form 19a, an internal two-electron transfer occurs from the π-system to the
S−S anti-bond, causing complete cleavage of the S-S bond and rotation of the
bipyridinium rings about the C4-C4’ bond. The internal electron transfer and concomitant
geometric rearrangement yields a HOMO almost entirely sulfur in character leaving the
aromatic bipyridinium rings free to accept additional electrons.
108
100
+
N
+
N
S
Current / A
50
S
3
2+
0
-50
-100
0.0
-0.5
-1.0
-1.5
-2.0
-2.5
+
E vs. Fc /Fc / V
Figure 5.6 Cyclic voltammogram of 0.73 mM 19a2+ in 0.10 M Bu4NPF6/acetonitrile at
298 K. Glassy carbon working electrode. Scan rate: 1.00 V/s. Line:
Background−corrected experimental cyclic voltammogram. Symbols: simulation for four
successive reversible electron−transfer reactions. The reactions were treated as reversible
(ks−values set at 0.3 cm/s). Standard potentials: E1 = −0.49; E2 = −0.60; E3 = −1.79;
E4 = −2.05 V.Dall species = 1.9  10−5 cm2/s.
109
The separated two−step reduction of 19a to 19a2- that involves addition of electrons to
the newly vacated π bipyridinium orbitals, and DFT computations show that they are
comprised of much the same character as those involved in the reduction from 19a2+ to
19a. However, with the S-S antibond already being populated and no internal electron
transfer possible, there is no major geometric rearrangement upon reduction.
On the return wave, 19a2- undergoes two one-electron oxidations with removal of the
electrons coming from the aromatic system. Oxidation of 19a then proceeds by the same
mechanism as 172- → 17 as shown in Scheme 5.2, but the driving force to go from the
rotated 19a+ to the non-rotated 19a+ with an intact S−S bond has been increased as can be
seen in Figure 5.7. Previously in 17− the geometric rearrangement was exergonic by 0.51
eV but in 19a+ it has increased to approximately 0.88 eV. This energy difference is
believed to cause the difference in kinetics and thus electrochemical reversibility between
the two molecules.
Thus, through the cyclic voltammetry studies coupled with DFT computations we
propose that the mechanism for 192+ involves two single electron transfer events into the
bipyridinium ring, followed by internal electron transfer from the π system to the S-S
antibond causing cleavage of the S-S bond and rotation about the C4-C4’ bond. This sets
the stage for 19 to accept two additional electrons into the bipyridinium rings. Once 192has been formed the return anodic waves liberates two electrons from the bipyridium
rings to form 19. From here on the oxidation pathway is the same as that for 172- as
shown in Scheme 5.2.
110
Figure 5.7 Energy rotation profile, with respect to the S-C4-C4’-S’ torsion angle for
19a2+ (top trace), 19a+ (second from top), 19a (third from top) 19a- (second from
bottom), and 19a2- (bottom trace).
111
In spite of the reversible processes seen in cyclic voltammetry, some slow chemical
reactions occur with longer timescale studies. Controlled potential electrolysis was
conducted on 19a2+ with the aim of generating a solution of 19a+. More than one Faraday
of charge was required per mole of the disulfide to achieve complete electrolysis. When
electrolysis was interrupted after passage of one Faraday per mole, voltammetric
investigation of the blue electrolysis solution was not consistent with the presence of the
radical cation, 19a+, but indicated the presence of other unknown products.
Furthermore, when 18+ was investigated electrochemically it was not well behaved,
due to precipitation onto the electrode surface and additional redox chemistry due to the
deposited species. Attempts were made to study 18+ by using samples from multiple
synthesis attempts, with exchange of the counterion (OTf
-
instead of PF6-) as well as
intense polishing between successive scans. Unfortunately, anomalous peaks were still
present and consistent results between scans of the same solution could not be obtained.
Consequently additional studies on 18+ were halted. Figure 5.9 shows preliminary results
for this compound.
112
Figure 5.8 LUMO of 19a2+ (top), SOMO of 19a+ (middle), and HOMO of 19a (bottom).
113
Figure 5.9 Cyclic voltammogram of 0.29 mM 18+ in 0.10 M n-Bu4NPF6/acetonitrile at
298 K. Glassy carbon working electrode. Scan rate: 1.0 V/s. The vertical bars at top are
the first, second, and third reduction potentials from the DFT computations. Note the
potential inversion of the first two reduction potentials.
114
Conclusions
In summary, we see considerably different behavior upon reduction of disulfides 17,
18+ and 19a2+, and the empirical electrochemical results are modeled and explained well
with simulations and DFT computations. For 17 two electrons are added to the disulfide
moiety followed by rapid rupture of the S−S bond forming the ring−opened dithiolate.
For 19a2+, two one−electron reductions of the bipyridinium ring cause an intramolecular
charge transfer to open the disulfide ring. After transfer of the electrons to the dithiolate,
the bipyridinium ring can once again be twice reduced. DFT calculations show that none
of the annulated compounds have a barrier to cleavage of the S−S bond upon reduction
by two electrons; however, in the return direction, barriers are present for ring closure.
The difference in the observance of potential inversion between 17 and 192+ result in
changes in the anodic wave of the cyclic voltammogram causing 19a+ to be reversible
while 17− is not. As one might expect, methylation of the bipyridine nitrogen atoms
increases π –character of the LUMO and moves the reduction potential to more positive
potentials.
The dominant factor in the chemistry of these disulfide radical anions is the ready
ability to accept a second electron, for which it is prepared by the large increases in the
S–S bond distance and the large reorganization energy that occurs with the first electron
reduction. Addition of the second electron occurs concomitantly with bond cleavage, that
is, it is a concerted dissociative electron transfer. Even in the gas phase the dianion
spontaneously separates the charged fragments to reduce the electron-electron repulsion.
These principles are likely to apply with appropriate adjustments to the reduction
chemistry of other systems with disulfide functionalities.
115
Experimental
Synthesis and characterization of 17−19a2+ were performed by other students
within the same collaboration and will be reported elsewhere, but involve the same
general procedures as is already present in the literature.115 Sources and treatment of
solvent and electrolyte were the same as reported earlier.46 Electrochemical procedures
including the determination and compensation of solution resistance have also been
reported.46 The working electrode was a 0.3−cm diameter glassy carbon disk whose area
was calibrated by studies of the oxidation of ferrocene in acetonitrile at 298 K using 2.5 
10−5 cm2/s as the diffusion coefficient of ferrocene.120 The laboratory reference electrode
was a silver wire in contact with 0.010 M AgNO3 in acetonitrile which also contained
0.10 M Bu4NPF6. Periodically the reversible potential for the ferrocenium/ferrocene
couple, Fc+/Fc, in acetonitrile was determined with respect to the silver reference
electrode, which allowed for post facto expression of all reported potentials versus
ferrocene. Simulations were carried out using DigiElch48 (http://www.elchsoft.com) with
either planar or disk geometry for the working electrode.
All computations were performed using the Amsterdam Density Functional Theory
program version adf2013.01.43,50,121 Computations were carried out with the PBE
functional122 (PBE-D3) with a d-ζ valence for hydrogen atoms and triple-ζ plus
polarization (TZP) for all other atoms. Relativistic effects are included by the zero order
regular approximation
123
(ZORA). Solvation free energies are estimated by the
conductor-like screening model124 (COSMO) of solvation using default parameters.
Additionally all geometries were optimized in solvent. Figures of the optimized
116
geometries and molecular orbitals were created with program Visual Molecular
Dynamics.55
117
Chapter 6 Asymmetrical Diiron Bisthiolate Complexes, [(–SR)(–
SR)Fe2(CO)6], for Electrocatalytic Proton Reduction from Weak
Acid
Introduction
Hydrogen is a clean fuel which can be used efficiently in fuel cells or combusted
without generation of greenhouse gases. This can serve to reduce demand on fossil fuels
which is an unsustainable and carbon based energy source. Additionally, as CO2
sequestration technology matures abundant supplies of H2 will be necessary to capture
CO2 and transform it into methanol or another valuable chemical feedstock. Currently the
primary source of molecular hydrogen is natural gas, which still relies on carbon based
byproducts. Ideally molecular hydrogen could be obtained from the splitting of water.
Currently the most energy efficient catalyst for this purpose is platinum. The use of
platinum is not practical due to the cost and limited natural abundance.125 In nature,
molecular hydrogen is produced as a fuel source by the hydrogenase class of enzymes
which catalyze the reversible reduction of protons at near thermodynamic potentials.126
The [FeFe] variety of hydrogenase has been shown to have the greatest rate of molecular
hydrogen production, and is comprised of Earth abundant and inexpensive elements.11
Unfortunately, the use of isolated enzymes as catalysts on a large scale is not practical
due to a lack of photo and thermal stability, low natural abundance and the difficulty of
the anaerobic purification process. Thus, investigation of complexes which are similar to
the active site is a prudent approach to developing efficient and inexpensive
electrocatalysts for hydrogen production.
The structure of the [FeFe]–hydrogenase active site, called the H–cluster consists of a
[4Fe4S] cubane linked to a [2Fe2S] cluster through a cysteine residue. The [4Fe4S]
cluster acts as a redox center for shuttling electrons to the [2Fe2S] active site where the
118
reversible proton reduction takes place. Synthesis and catalytic use of the active site has
been attempted13,127 but is limited due to high sensitivity of cyanide and the cubane
cluster, and a loss of performance outside of the enzyme framework.
Much effort has been devoted on the development of an efficient synthetic catalyst by
mimicking the [FeFe]–hydrogenase active site with a simpler structural motif while
retaining highly efficient catalytic function. The active–site mimics of the formula [(–
SR)2Fe2(CO)6]4 are widely studied such as [(–benzenedithiolato)2Fe2(CO)6],15,16 [(–
ethylenedithiolato)2Fe2(CO)6,128 and [(–1,3–propanedithiolato)2Fe2(CO)6]129-131 for all
of which proposed mechanisms exist in the literature.
Dye–sensitized solar cells utilizing thiophene and oligothiophenes have received much
attention due to their optoelectronic properties.132-134 Inspired by this, our group
decorated the [(–S2)Fe2(CO)6] core with thiophene to act as a model for future systems
with oligiothiophenes acting as chromophores to initiate light–driven hydrogen
production.135,136 In this work, a new asymmetric bis–(–thiolato) diiron complex, [(–
SMe)(–STh)Fe2(CO)6] (23) and an analogous complex [(–SMe)(–SPh)Fe2(CO)6]
(24), shown in Scheme 1, are studied for use as catalysts. Complex 23 was designed as a
first step toward (µ–oligothienylthiolate)[Fe(CO)3]2 for photo–assisted H2 production.
119
Results and Discussions
Synthesis and Structural Analysis
The asymmetric bis–thiolate diiron complexes 23 and 24 were synthesized by
modification of literature procedure as detailed in Scheme 6.1,137 and was primarily
carried out by a visiting student in our lab, Orrasa In-noi, and consequently will not be
discussed in detail here. A manuscript containing the experimental details is being
prepared.
The anti–isomeric structures of 23 and 24 were identified by NMR spectroscopy and
single crystal X–ray diffraction. The 1H NMR spectrum of 24 in CDCl3 solution shows
two methyl signals at 1.72 and 2.06 ppm corresponding to the axial and equatorial
positions of the methyl group, respectively. This is in agreement with what was originally
found.137 Compound 23 showed multiple isomers in the NMR spectrum as well and was
found to have an isomer ratio for 23a:23b of approximately 2:1 upon work-up.
Attempts to study the equilibration between isomers of 24 by variable temperature
NMR were not definitive. While performing variable temperature NMR experiments
precipitation occurred at 85C due to decomposition and consequently observation of
methyl group interconversion between axial and equatorial was not possible. This agrees
with a report on [(–SMe)2Fe2(CO)6] which shows no evidence of axial–equatorial ligand
exchange.138
120
Scheme 6.1 (a) THF, 2-ThMgBr, -78 ºC, 30 min; (b) CH3I, -78 ºC, 30 min; ambient
temperature, 20 hr.
121
24a
24b
23a
Figure 6.1 Thermal ellipsoid plots of molecular structures of 23a, 24a, b, and an overlay
of the crystal structures of 23a and, 24a, all viewed along the Fe-Fe bond. Display of 24a
and 24b is half the asymmetric unit. Thermal ellipsoid plots at 50% probability,
Hydrogen atoms have been omitted for clarity.
122
Recrystallization from saturated pentane solutions kept at -18 C for 4–5 days yielded
suitable crystals for single crystal X-ray diffraction. X–ray crystallography revealed the
crystal structure of 23 to exist solely as a single anti–isomer with the methyl group in the
equatorial position and the thienyl substituent in the axial position (i.e. 23b was not
present in the crystal). Crystals of 24 were grown from a mixture of two anti–isomers, in
which the methyl group was in either the equatorial (24a) or axial position (24b), in
saturated pentane solution and the two isomers cocrystallized in a 50:50 ratio. The
molecular structures of the compounds are displayed in Figure 6.1.
Molecular structures of 23–24 contain two Fe(CO)3 subunits which are bridged by two
different thiolate ligands, resulting in a [2Fe2S] core in the familiar butterfly framework.
The Fe-Fe bond lengths of 2.5116(6)-2.5215(3) Å, fall in the normal Fe–Fe bond length
for [2Fe2S] hydrogenase mimics containing six CO and two thiolate ligands (2.49–2.57
Å)139 providing a pseudo–octahedral environment around each iron center.
Examination of key distances about the 2Fe2S core shows many of the bonds, such as
the Fe-Fe bond to be statistically identical in the solid state of 23a (2.5215(3) Å), and
24a (2.5116(6) Å). The most notable difference between 23a and 24a in the solid state
structures is the orientation of the aromatic ring in relation to the 2Fe2S cubane core. In
the crystal structure when defining two planes, one defined by all non-hydrogen atoms in
the aromatic ring, and the other defined by the two Fe atoms and two apical CO’s, the
twist between the two planes in is 18 ° greater for 24a than it is for 23a. In fact in 23a the
twist is small enough to be negligible. This leads to the question of whether this change
in the angle of the aromatic plane is due to intramolecular electronic effects or if it is due
to packing interactions. Analysis of both structures shows no close packing interactions
123
within van der Waals radii for either of the aromatic rings in the axial position.
Additionally this increased twist angle is reproduced in gas phase calculations albeit at a
reduced amount of an 11° difference between 23a and 24a. Examination of the HOMO –
8 of each complex Figure 6.4 shows an increased degree of double bond character
between the 2Fe2S unit and the aromatic ring for 23a. A formal π bond between these
two moieties would suggest increased communication and delocalization between the
two. A possible point of interest with regard to future applications with oligiothiophenes
as photocatalysts.
124
LUMO
HOMO
Homo – 1
Homo – 2
Figure 6.2 LUMO through HOMO – 2 of 23a and 24a.
125
Homo – 3
Homo – 4
Homo – 5
Homo – 6
Figure 6.3 HOMO – 3 through HOMO – 6 of 23a and 24a.
126
Homo – 7
Homo – 8
Homo – 9
Homo – 10
Figure 6.4 HOMO – 7 through HOMO – 10 of 23a and 24a.
127
60
40
20
I /A
0
-20
-40
-60
-80
1.5
1
0.5
0
E/V vs.
-0.5
-1
-1.5
Fc+/Fc
Figure 6.5 Anodic and cathodic scans of 23a (solid line) and 24a (dashed line) in
CH3CN under N2 atmosphere. Arrows indicate the direction of scans.
-2
128
Electrochemistry
Figure 6.5 shows the cyclic voltammograms of 23a and 24a in acetonitrile. Both 23a
and 24a display irreversible reduction and oxidation processes. The reductions occur at a
peak potential of -1.46 and -1.53 V vs. Fc+/Fc respectively. A 70 mV positive shift in
reduction potential is observed from 23a to 24a. This is somewhat counter to what one
might expect, typically a thiophene moiety is considered to be more electron rich than a
corresponding phenyl moiety, and consequently should be more difficult to reduce. The
oxidation potential of 23a shifts 60 mV cathodic than that of 24a; i.e. in the opposite
direction as one would expect from the anodic shift of the reduction potential. If the
HOMO and LUMO have similar compositions the oxidation and reduction potentials
should track in the same direction. Taken together the reduction and oxidation potentials
of 23a are spaced 130 mV closer than in 24a. This suggests that either the HOMO or the
LUMO of 23a and 24a have different atomic orbital contributions. Additionally the νCO
region of the infrared spectrum shows that the stretching frequencies of 23a are slightly
higher in frequency than those of 23a Figure 6.6. This suggests that there is indeed
increased electron density at the metal centers and increased π-backbonding to the CO’s
in 23a. Thus the thiophene moiety is less electron donating than phenyl into the occupied
metal orbitals decreasing νCO stretching frequencies and making oxidation easier and
more electron withdrawing than phenyl to the unoccupied metal orbitals making
reduction more facile.
When examining the cathodic chemistry of 23a and 24a, the return anodic scan shows a
small oxidation peak at about -0.5 V (Figure 6.5) corresponding to the oxidation of
products formed upon reduction of the neutral complex. This oxidation peak is similar to
the peak found in reported complexes; [(–SPh)2Fe2(CO)6],140 [(–S–2–RCONHC6H4)2
129
Fe2(CO)6] (R= CH3, CF3, C6H5, 4–FC6H4), [(–RS)2Fe2(CO)6] (R = Me, Et)141 and [(–
1,3–propanedithiolato)2Fe2(CO)6]142 which are assigned to the oxidation of an
unidentified electro-active fragment or chemical transformation of reduced intermediate
to other stable species which are more difficult to oxidized.
While not tied to the reduction on the electrochemical time scale, the oxidation at ~ -0.5
V after reduction plays an important role in recovery of the neutral species. Performing 3segment cyclic voltammetry experiments showed that if the return oxidation at -0.5 V is
scanned through then most of the cathodic current can be recovered on the third segment
returning through -1.59 V (-1.09 V → -1.59 V → -1.09 V → -1.59 V). However if the
second segment is stopped short of the oxidation at -0.5 V then appreciable loss of
cathodic current occurs on the third segment. To insure that the increased current was not
from the increased potential window and concomitant increased diffusion due to
increased experimental time, vertex delays were employed. In order to achieve full
recovery of cathodic current without scanning past the oxidation at -0.5 V a delay of at
least 30 s was necessary. Comparatively the time elapsed when running a three segment
scan of the form -1.09 V → -1.59 V → 1.09 V → -1.59 V at 100 mV/s is only 22 s as
and can recover all of the current. Recovery of current remains even at faster scan rates
(200 mV/s) which allows only 11s to elapse. This suggests that the reduced species may
undergo a chemical transformation to a stabilized form of the reduced species such that it
is more difficult to oxidize on the return scan, but remains a chemically reversible
process.
130
Figure 6.6 IR spectra in CO region of 23a (solid line) and 24a (dashed line) in mineral
oil.
131
2200
2100
2000
1900
1800
1700
1600
Frequency (cm-1)
Figure 6.7 IR-SEC spectra obtained by reduction of 23a (top) at -1.4 V and 24a
(bottom) at -1.2 V Vs. a pseudo silver reference in an Ar saturated solution of CH3CN.
The solid line indicates the spectrum post-electrolysis and dashed line indicates the
spectrum pre-electrolysis
132
In order to obtain more information on the structure of the reduced species, a bulk
electrolysis experiment was performed on 24a in a CO saturated solution of THF with
Bu4NPF6 as supporting electrolyte. The IR spectrum of the electrolyte solution before
bulk electrolysis, showed three n(CO) IR bands at 2071, 2034, and 1991 cm-1. The
potential in the bulk electrolysis experiment was held at -1.6 V for 10 min to produce a
large amount of reduced product. The new IR spectrum shows three new bands at lower
frequencies, 2016, 1967, and 1774 cm-1 related to the reduction product. The new n(CO)
IR band at 1774 cm-1 corresponds to bridging CO coordination which is not possible with
a monoiron species suggesting the molecule is either staying intact or dimerizing.
Additionally this is consistent with what has been seen for similar Fe2S2(CO)6 species
that have the S atoms bridged.15,16
IR Spectroelectrochemistry of 23a and 24a in acetonitrile under an Ar atmosphere, as
shown in Figure 6.7, was also performed to obtain information on the reduced species
with a shorter lifetime than the bulk electrolysis experiment. With no initial potential,
both compounds show three IR absorption bands at approximately 2000, 2040 and 2075
cm-1 corresponding to three terminal CO stretching frequencies. A linear sweep through
the reduction, results in four new peaks from 2028 to 1925 cm-1 and a single broad low
wavenumber ν(CO) band for both the reduced forms present at about 1715 cm-1. These
infrared spectra in combination with the necessity of passing through the reduction at
~0.5 V in order to recover current, has led to the conclusion that while not reversible, the
reduced species does not break apart to monoiron compounds after cleavage of the Fe-Fe
bond. Also of note is that while 23a has stretching frequencies which are  2 cm-1 higher
in the neutral complex, while the anions have statistically identical stretching frequencies.
133
Gas-phase Photoelectron and UV-Vis Spectroscopy
Photoelectron spectroscopy is a powerful tool for elucidating the electronic structure of a
molecule and the character of its molecular orbitals. More specifically intensity changes
of an individual peak between He I and He II spectra gives insight into the changes in
cross section for ionization of the responsible orbital which is dependent upon the atomic
orbital contributions. Examination of Figure 6.8 shows that there is minimal drop off in
intensity transitioning from He I to He II for the leading edge of the spectra for
compound 24a suggesting the HOMO is primarily Fe-Fe bonding in nature. This
however is not the case for compound 23a. The drop in intensity on the leading edge
leads to the conclusion that there is significant sulfur or carbon (i.e. ligand) mixing into
the HOMO of 23a. DFT computations support this as shown in Figure 6.2-6.4.
The UV–Vis absorption of 23a and 24a in hexane solution shows broad absorption
bands at about 330 nm (Figure 6.9). Complex 23a (13,300 M-1cm-1) exhibits a slightly
larger  for λmax than 24a (11,400 M-1cm-1). These  values are consistent with weak
charge transfer showing considerable ligand character in one of the orbitals involved in
the excitation. The larger  for 23a is in agreement with other data presented herein
which suggests that the HOMO of 23a has more aromatic character than 24a.
134
Figure 6.8 He I (solid black) and He II (dashed red) ultra-violet photoelectron spectra of
23a (top) and 24a (bottom).
135
Figure 6.9 UV-Vis absorbtion spectra of 0.2 mM 23a (solid line) and 24a (dashed line)
in hexane solution.
136
Electrocatalytic Proton Reduction of Weak Acid
Electrocatalytic production of hydrogen by 23a and 24a was tested with acetic acid
(HOAc) in acetonitrile. Voltammograms are shown in Figure 6.10. After addition of acid
a second peak at -1.8 to -1.9 V appears and the current is increased with increasing acid
concentration. This behavior is analogous to that of known proton reduction catalysts
such as (–1,2–benzenedithiolato)Fe2(CO)6 (4)47 and (–1,3–propanedithiolato)Fe2(CO)6
(PDT).143 Thus, the first reduction potential is related to primary reduction of neutral
catalyst precursors to active states and the second potentials correspond to hydrogen
production. The overpotentials of proton reduction from HOAc of 23a (-0.36 V) and 24a
(-0.42 V) are low compared to 4 (-0.65 V) and PDT (-0.89 V).4
A new behavior was observed in voltammogram of 23a. At added acid concentration of
2 mM to 20 mM, the current of the catalytic peak of 23a is higher than those obtained
from 24a until at 50 mM HOAc the catalytic current of 23a no longer increases and peak
potential shifts about 0.1 V in the negative direction. This observation suggests that either
a different reduction process may occur at decreased pH, or free acetate in solution may
be decomposing the catalyst. Another observation in the presence of acid is a positive
shift in the first reduction potential. The presence of HOAc results in a minimal shift (30–
40 mV) of the primary reduction to more positive potentials suggesting precoordination
of the acid and catalyst and/or rapid protonation of the reduced catalyst. Electrocatalytic
proton reductions of 23a in presence of acids with differing pKa’s, phenol (29.1), 4–
bromophenol (25.5), and cyanoacetic acid (18.0),144 were investigated. In the presence of
the weakest acid, 4-bromophenol, no shift of the initial reduction potential and no
catalytic reduction is observed with sequential additions of the phenol. With stronger
137
acids the extent of the positive shift of the reduction potential and catalytic current is
increased proportional to the acid strength.
Repeating cathodic scans in the presence of acetic acid shows the peak at -2.2 V to
progressivly shift in the positive direction to about -2.0 V. This is interpreted as an
electro-active reduction product precipitating onto the electrode surface, and
consequently all potentials reported herein are for a freshly polished electrode. The
positive shift of catalytic potential about 0.2 V shows potential for future applications
such as heterogeneous catalysis in aqueous media.
Under a CO atmosphere, the catalytic reduction peak at -1.8 to -1.9 V observed under
N2 is greatly diminished. Either additional CO inhibits formation of the catalytically
active species or CO reacts with the catalyst to effectively remove it from the cycle.
Under CO with and without added acetic acid no shift in reduction potential at -2.2 V
suggests that no electrode coating occurs in a CO–saturated solution. Further studies of
the catalytic mechanism of the proton reduction process by 23a and 24a are in progress.
138
390
[HOAc]/
mM
50
290
190
I / A
20
10
90
5
2
0
-10
-1.3
-1.6
-1.9
E / V vs.
Fc+/
-2.2
Fc
390
[HOAc]/
mM
50
290
190
I / A
20
10
90
5
2
0
-10
-1.4
-1.7
E / V vs.
Fc+/
-2
Fc
Figure 6.10 Voltammograms of ca. 1 mM 23a (top) and 24a (bottom) in 0.10 M
Bu4NPF6/CH3CN at 0.10 V/s in the absence and presence of various concentrations of
HOAc. Red arrow indicates initial direction of potential sweep. Return waves have been
omitted for clarity
139
Conclusions
A new [FeFe]–hydrogenase active–site inspired complex, [(–methylthiolato)(–2thiophenethiolato)Fe2(CO)6], was successfully synthesized from the Grignard cleavage of
(–S)2Fe2(CO)6 followed by methylation. The major product was confirmed by IR and
NMR spectroscopy and X–ray crystallography to be the anti–isomer with an equatorial
methyl group and axial thienyl group. Additionally this complex has been compared to
the similar known complex [(–methylthiolato)(–phenylthiolato)Fe2(CO)6]. Structurally
the 2Fe2S core of these molecules is changed minimally, but the thiophene substituent
electronically
modulates
the
complex
showing
increased
non-innocence
and
delocalization. Both complexes are determined to remain intact after reduction, and while
the reduction is electrochemically irreversible the electrochemical reduction is thought to
be a chemically reversible reduction. Both complexes show electrocatalytic reduction of
protons from weak acid to molecular hydrogen with a lower overpotential than many of
the known molecules of this class. Thiophene substitution reorders the molecular orbitals
to cause the HOMO of 24a to be heavily ligand based while the LUMO is metal based.
This observation opens the possibility of ligand to metal charge transfer which is a good
sign for light-driven hydrogen generation by [FeFe]–hydrogenase inspired catalysts
appended with oligiothiophene chromophores.
140
Experimental Section
General Procedures.
All reactions were performed under an Ar atmosphere using standard Schlenk line
techniques. Solvents were deoxygenated by sparging with Ar for 30 min prior to use.
Workup and chromatographic separations were carried out in air using silica gel columns
with pentane or hexanes as eluents. Product yields were calculated based on utilized [(–
S2) Fe2(CO)6]. Infrared spectra were collected in mineral oil or in hexane on round
sodium chloride cells using a Nicolet 380 FT–IR spectrophotometer. UV-Vis absorption
spectra were collected in 0.2 mM of the synthesized complexes in hexane solution with a
quartz cuvette with a path length of 1 cm using a Agilent 8453 UV-Vis spectrometer.
X–ray Structure Determination
Suitable red crystals of both complexes were obtained by recrystallization from
saturated pentane solutions kept at -18 C for several days. All X–ray crystallography
data was collected by the X–ray Diffraction Facility in the Department of Chemistry and
Biochemistry, The University of Arizona, and was collected using a Bruker–Apex II Duo
diffractometer equipped with a Mo K X–radiation source ( ̅ = 77.17.0 Å). The crystal
structure of [(–SMe)(–SPh)Fe2(CO)6] are solved by direct methods and the crystal
structure of [(–SMe)(–STh)Fe2(CO)6] are solved by Patterson methods and refined by
full–matrix least–squares refinement on F2 using eht SHELXTL software.145 All
nonhydrogen atoms were refined anisotropically and hydrogen atoms were placed in
idealized positions. Figures were generated using Olex2.146
141
Electrochemistry
Sources and treatment of solvent and electrolyte are the same as reported earlier.46
Electrochemical procedures including the determination and compensation of solution
resistance have also been reported.46 Cyclic voltammetry experiments were carried out
using an EG & G Princeton Applied Research potentiostat/Galvanostat model 273. A
standard 3–electrode system was utilized including a Ag/0.01 M AgNO3 in CH3CN
reference electrode, a glassy–carbon working electrode (GCE) with 3 mm diameter
determined area to 0.0878 cm2, and a Pt wire auxiliary electrode. The reference electrode
was corrected against a ~ 1.0 mM solution of ferrocene in acetonitrile and all potentials
are reported against Fc+/Fc.147 General conditions for the cyclic voltammetry experiments
included approximately 1 mM of each diiron compound in acetonitrile containing 0.1 M
n–Bu4NPF6 as the supporting electrolyte, a scan rate of 0.100 V/s, ambient temperature,
and a minimum 30 seconds between scans with stirring under N2 or CO. The electrode
was polished with 0.05 micron alumina in deionized water on a felt surface prior to each
experiment to prevent reduced catalyst from previous voltammetric experiments from
remaining on the working electrode surface. Reticulated vitreous carbon was used as
working electrode for bulk electrolysis experiments due to its large surface area in order
to increase the rate of electrolysis.
IR spectroelectrochemistry
IR-SEC experiments were conducted in a cell of similar design to what exists in the
literature.57 Solutions were made by sparging a solution of 0.2 M TBAH electrolyte in
CH3CN with Ar for 20 minutes and then dissolving approximately 1 mg of compound per
1 mL of solution. A Gamry ref 3000 potentiostat was used to perform a linear sweep
voltammogram past the point of reduction. An IR spectrum was then collected on a
142
Thermo Nicolet Avatar ESP 380 FT-IR spectromenter. The spectrum of CH3CN and
TBAH was then manually subtracted.
Gas–phase UV Photoelectron Spectroscopy
Photoelectron spectra were recorded using an instrument that features a 36 cm
hemispherical analyzer (McPherson) with custom–designed photon source, sample cells,
detection and control electronics. Spectrum calibration was obtained with Ar, and data
fitting by WinFp v22.09 program.148 All samples sublimed at 70–80 C. The contour of
ionization intensity obtained at 21.2 eV (He I).
Density Functional Theory Calculations
DFT calculations were performed using Amsterdam Density Functional (ADF)
software version 2010.02.149 Geometry optimizations and frequency calculations were
carried out in the gas–phase using local density approximation (LDA)150 and the Vosko–
Wilk–Nusair (VWN) functional151 with the Stoll correction implemented.54 A
polarization function of triple–zeta with one polarization function (TZP) and relativistic
corrections zeroth–order relativistic approximation (ZORA),43 available in the ADF
package were used in all calculations. Molecular orbital plots were created using the
visualization program VMD55 with a surface cutoff value of 0.05.
143
Chapter 7 Conclusions and Future Directions
By combining spectroscopic techniques with voltammetric studies and insights gained
by density functional theory computations considerable insight has been gained into the
redox mechanisms of various sulfur and iron containing compounds. The compounds
studied focused on answering one of two questions; the manner in which the S-S bonds
cleave upon reduction, and understanding how changes in ligand architecture modulates
the electronic structure and redox behavior of 2Fe2S compounds. The bulk of this
dissertation focused on the latter of these two; with the specific aim of how to better
understand the effects of ligand architecture on electro-catalytic hydrogen production by
the complexes studied herein. Special attention is given to understanding the interplay of
the catalyst core with “non-innocent” redox active and chromophoric ligands to aid future
developments in photo-catalytic pathways for molecular hydrogen production. This is the
focus of the research found ino Chapters 3, 4, and 6.
Chapter 3 focuses on a series of (μ-dithiolato)Fe2(CO)6 compounds in which the
dithiolato sulfur atoms are bridged by 5-substituted 1,4-benzoquinones, 1,4naphthoquinone, or 1,4-anthraquinone. These complexes were studied by cyclic
voltammetry, IR spectroscopy, EPR spectroscopy, and DFT computations. These
complexes were found to show a different reduction pathway than many other enedithiolates of this class; they form semi-stable radical anions rather than undergoing a
significant geometric rearrangement resulting in potential inversion. Thus, the focus of
the study was shifted to using these complexes to understand the reduction behavior and
electron delocalization of the once reduced species of diiron-enedithiolates via EPR
spectroscopy. Attachment of the S2Fe2(CO)6 to the quinone showed a shift in reduction
144
potential approximately equivalent to attachment of three chlorine atoms to free
benzoquinone. Additionally by matching the EPR hyperfine values, IR stretching
frequencies and structural parameters to computed values with multiple computational
packages, the computational model employed was validated for this type of complex. The
computational results in combination with the percent decrease of aH from the free
quinones to the complexes, showed that there is extensive electronic communication and
spin delocalization in the radical anion of these complexes, with the ability to tune the
spin density present on iron from 17% up to 23% by varying the functionality of the 1,4quinone from 2-chloro-1,4-benzoquinone to 1,4-naphthoquinone. This is potentially
useful for future applications of trying to tune spin and electron density to change
protonation sites.
While Chapter 3 can be considered a completed project there are insights gleaned
which could be built upon in future projects. The major drawback of the present system
was that the oxygen atoms proved to be the site of protonation rather than acting as a
proton transfer site to the Fe-Fe bond. Designing a system that contains a similar type
internal base which will encourage protons to be in the proximity of the active site for
catalysis, but will concentrate a greater percentage of electron density on the Fe centers to
encourage protonation of the Fe-Fe bond by the proton transport base could still be
promising. This could comprise something as simple as replacing the quinone with an
aromatic aniline dithiolate for which the pKa could be tuned by functionalization, or
variation of the solvent. From an electronegativity standpoint the nitrogen atom should
withdraw less electron density and allow more to reside on the metal centers.
Additionally, with ammonium type acids being comprised of a neutral base and cationic
145
acid, the pKa of such a system varies greatly with solvent polarity, giving a variable that
could be optimized for a mechanistic study on compounds containing an internal base.
With substitution of the quinone by an aniline based ligand the mechanism or reduction
would more than likely be altered to something more similar to that of benzcat15 rather
than the two sequential one-electron reductions of the present system, but that is not of
concern if the greater goal of facile hydrogen production is ultimately met.
Chapter
4
studied
two
classic
functional
hydrogenase
mimics,
(–1,2–
benzenedithiolato)Fe2(CO)6 and (–1,3–propanedithiolato)Fe2(CO)6 which had been
appended with a chromophore, 2-phenyl azo pyridine (PAP), in place of two of the
carbonyl ligands on one of the Fe atoms. The synthesized complexes displayed an intense
blue color whereas all other complexes of this type studied by our research group have an
orange or red color. Thus, electronic interplay between the 2Fe2S core and the PAP
chromophore was studied via UV-Vis spectroscopy. The combination of a high ϵ and shift
in absorption demonstrated that a charge transfer was present, but the classic test of
monitoring λmax while changing solvent polarity did not significantly change λmax.
DFT and TD-DFT computations were successfully used to model this system. By
implementing TD-DFT computations we found that the HOMO through HOMO -6 were
composed almost entirely of 2Fe2S character based orbitals while the LUMO was
comprised of approximately 65% PAP character and that this was the most probable
transition. The presence of charge transfer was further demonstrated by computing the
difference in energy between the ground state and the excited states in a variety of
solvents of varying dielectric constant. The energy difference between the two electronic
configurations was found to vary minimally with solvent suggesting that the structure of
146
the molecule made it difficult for solvent to access the charged portions of the molecule
to alter the degree of solvent stabilization between the ground and excited states. Thus for
this class of compounds monitoring λmax with varying solvent polarities is a poor test for
the presence of charge transfer.
While the compounds in Chapter 4 were a nice proof of concept and allowed us to
study the way the molecular orbitals interact in these systems they ultimately proved to
be unstable and impractical for catalysis due to degradation. Identification of the
degradation products could be investigated in order to study how to modify the catalysis
for greater durability, but ultimately it is my opinion, the best approach would be to scrap
coupling the 2Fe2S core with PAP, and either choose a different chromophore to replace
the CO ligands or change the site of chromophore attachment to the bridging thiolates,
which ultimately could lead to the ability of tuning/eliminating back electron transfer by
changing a linker between the chromophore and bridging sulfur atom.
Chapter 5 is slightly divergent from the general theme of hydrogen production,
but looks at a related fundamental question which is the mechanism for cleavage of
disulfide bonds. For this study a series of 4,4’-bipyridine 3,3’-disulfide complexes was
chosen; particular emphasis being placed whether the site for reduction is the
bipyridine/bipyridinium ring or the S-S, and if the S-S bond cleaves after a one electron
reduction or after a two electron reduction. This chapter successfully used cyclic
voltammetry coupled with DFT computations to show that for 4,4’-bipyridine-3,3’disulfide in acetonitrile reduction within the solvent window is limited to a two electron
process with substantial potential inversion. Upon reduction, electrons are inserted into
the S-S antibond, and while the disulfide bond lengthens with the first electron, cleavage
147
does not occur until after addition of a second electron. In the case in which both
bipyridinium rings have been methylated to afford an overall + 2 charge in the unreduced
molecule, it was found that the molecule undergoes four separated one electron processes
within the solvent window. The first two reductions occur into the bipyridinium ring and
then internal electron transfer of both electrons results in cleavage of the S-S bond with
minimal thermodynamic barriers after electron transfer. DFT computations show that the
remaining two reductions take place into essentially the same orbitals as initial insertion
of the first two reductions.
Chapter 5 has culminated to the point where it has been submitted as a paper to
JACS, but still retains areas in which it could be extended. Extending the study to include
a variety of different disulfide bonds could start to show trends for disulfide bonds as a
whole rather as well as different behavior of subclasses of disulfide bonds. Additionally
much of this chapter could be and currently is being extended through more complex
analysis of the computations similar to the treatment used by Savéant.152,153
Chapter 6 focuses on “open” systems in which the two thiolate sulfur atoms are
not bridged by an organic moiety. This project compares and contrasts of [(–SMe)(–
STh)Fe2(CO)6] and [(–SMe)(–SPh)Fe2(CO)6] with one another while keeping possible
substitution of oligiothiophenes (to act as a chromophore) of various lengths for the
thiophene thiolate in mind. This study showed that the reduction and oxidation potentials
of [(–SMe)(–STh)Fe2(CO)6] were spaced more closely together than those of [(–
SMe)(–SPh)Fe2(CO)6]. This combined with DFT and UPS showed that the HOMO of
[(–SMe)(–STh)Fe2(CO)6] was in large part ligand based, which is promising for the
possibility of future photo-catalytic systems. The electrochemical behavior of both
148
complexes was also examined in the presence of acetic acid and both complexes were
shown to have smaller overpotentials, albeit at lower rates of hydrogen production, than
both as (–1,2–benzenedithiolato)Fe2(CO)6 and (–1,3–propanedithiolato)Fe2(CO)6.
Both of which are classic hydrogenase mimics.
While Chapter 6 exists as a written paper (publication is uncertain while writing this),
the logical trajectory to follow on this project is clear: synthesize the oligiothiophene
version of various length oligiothiophene lengths and test them for photo-catalytic
production of hydrogen.
While this dissertation has a specific emphasis on studying hydrogen production
and manners in which the production can be increased in a cost effective manner, the true
focus of this dissertation for me is a better understanding about how the electronic
structure of iron and sulfur compounds are perturbed with different substitutions and how
that will affect their electrochemical reduction mechanisms and reactivity in various
oxidation states. By combining the use of various techniques ranging from cyclic
voltammetry, ultra-violet photoelectron spectroscopy, to DFT computations a more
complete understanding of how and why the complexes studied within this dissertation
was achieved than could be by any one technique alone. This multifaceted approached
has proved invaluable and shown the true merits of collaborative research.
149
References
References
(1) Peters, J. W.; Lanzilotta, W. N.; Lemon, B. J.; Seefeldt, L. C. Science 1998, 282,
1853.
(2) Nicolet, Y.; Piras, C.; Legrand, P.; Hatchikian, C.; Fontecilla-Camps, J. C. Struct.
Fold Des. 1999, 7, 13.
(3) Fontecilla-Camps, J. C.; Volbeda, A.; Cavazza, C.; Nicolet, Y. Chem. Rev. 2007, 107,
4273.
(4) Felton, G. A. N.; Mebi, C. A.; Petro, B. J.; Vannucci, A. K.; Evans, D. H.; Glass, R.
S.; Lichtenberger, D. L. J. Organomet. Chem. 2009, 694, 2681.
(5) Tschierlei, S.; Ott, S.; Lomoth, R. Energy Environ. Sci. 2011, 4, 2340.
(6) Tard, C.; Pickett, C. J. Chem. Rev. 2009, 109, 2245.
(7) Tran, P. D.; Barber, J. Phys. Chem. Chem. Phys. 2012, 14, 13772.
(8) Wang, M.; Chen, L.; Sun, L. Energy Environ. Sci. 2012, 5, 6763.
(9) Gloaguen, F.; Rauchfuss, T. B. Chem. Soc. Rev. 2009, 38, 100.
(10) Heinekey, D. M. J. Organomet. Chem. 2009, 694, 2671.
(11) Frey, M. ChemBioChem 2002, 3, 153.
(12) Razavet, M.; Borg, S. J.; George, S. J.; Best, S. P.; Fairhurst, S. A.; Pickett, C. J.
Chem. Commun. 2002, 7, 700.
(13) Tard, C.; Liu, X.; Ibrahim, S. K.; Bruschi, M.; De Gioia, L.; Davies, S. C.; Yang, X.;
Wang, L.; Sawers, G.; Pickett, C. J. Nature 2005, 433, 610.
(14) Wang, F.; Wang, W.; Wang, H.; Si, G.; Tung, C.; Wu, L. ACS Catal. 2012, 2, 407.
(15) Felton, G. A. N.; Vannucci, A. K.; Chen, J.; Lockett, L. T.; Okumura, N.; Petro, B.
J.; Zakai, U. I.; Evans, D. H.; Glass, R. S.; Lichtenberger, D. L. J. Am. Chem. Soc.
2007, 129, 12521.
(16) Wright, R. J.; Zhang, W.; Yang, X.; Fasulo, M.; Tilley, T. D. Dalton Trans. 2012,
41, 73.
150
(17) Lyon, E. J.; Georgakaki, I. P.; Rabenspies, J. H.; Darensbourg, M. Y. Angew. Chem.
Int. Ed. 1999, 38, 3178.
(18) Borg, S. J.; Bondin, M. I.; Best, S. P.; Razavet, M.; Liu, X.; Pickett, C. J. Biochem.
Soc. Trans. 2005, 33, 3.
(19) de Carcer, I. A.; DiPasquale, A.; Rheingold, A. L.; Heinekey, D. M. Inorg. Chem.
2006, 45, 8000.
(20) Wang, Z.; Liu, J.; He, C.; Jiang, S.; Åkermark, B.; Sun, L. J. Organomet. Chem.
2007, 692, 5501.
(21) Li, H.; Rauchfuss, T. B. J. Am. Chem. Soc. 2002, 124, 726.
(22) Wang, W.; Wang, H.; Si, G.; Tung, C.; Wu, L. Dalton Trans. 2009, 15, 2712.
(23) Wang, Z.; Liu, J.; He, C.; Jiang, S.; Akermark, B.; Sun, L. Inorg. Chim. Acta 2007,
360, 2411.
(24) Ezzaher, S.; Capon, J.; Gloaguen, F.; Petillon, F. Y.; Schollhammer, P.; Talarmin, J.
Inorg. Chem. 2007, 46, 9863.
(25) Gao, W.; Ekstroem, J.; Liu, J.; Chen, C.; Eriksson, L.; Weng, L.; Akermark, B.; Sun,
L. Inorg. Chem. 2007, 46, 1981.
(26) Si, Y.; Ma, C.; Hu, M.; Chen, H.; Chen, C.; Liu, Q. New J. Chem. 2007, 31, 1448.
(27) Gao, W.; Liu, J.; Ma, C.; Weng, L.; Jin, K.; Chen, C.; Åkermark, B.; Sun, L. Inorg.
Chim. Acta 2006, 359, 1071.
(28) Jiang, S.; Liu, J.; Shi, Y.; Wang, Z.; Akermark, B.; Sun, L. Polyhedron, 2007, 26,
1499.
(29) Jiang, S.; Liu, J.; Shi, Y.; Wang, Z.; Akermark, B.; Sun, L. Dalton Trans. 2007, ,
896.
(30) Hou, J.; Peng, X.; Liu, J.; Gao, Y.; Zhao, X.; Gao, S.; Han, K. Eur. J. Inorg. Chem.
2006, 22, 4679.
(31) Song, L. C.; Yang, Z. Y.; Bian, H. Z.; Liu, Y.; Wang, H. T.; Liu, X. F.; Hu, Q. M.
Organometallics 2005, 24, 6126.
(32) Eisenberg, R.; Gray, H. B. Inorg. Chem. 2011, 50, 9741.
(33) Sproules, S.; Wieghardt, K. Coord. Chem. Rev. 2011, 255, 837.
151
(34) Eisenberg, R. Coord. Chem. Rev. 2011, 255, 825.
(35) Joshi, H. K.; Enemark, J. H. J. Am. Chem. Soc. 2004, 126, 11784.
(36) Joshi, H. K.; Inscore, F. E.; Schirlin, J. T.; Dhawan, I. K.; Carducci, M. D.; Bill, T.
G.; Enemark, J. H. Inorg. Chim. Acta 2002, 337, 275.
(37) Cooney, J. J. A.; Cranswick, M. A.; Gruhn, N. E.; Joshi, H. K.; Enemark, J. H.
Inorg. Chem. 2004, 43, 8110.
(38) Cranswick, M. A.; Dawson, A.; Cooney, J.,Jon A.; Gruhn, N. E.; Lichtenberger, D.
L.; Enemark, J. H. Inorg. Chem. 2007, 46, 10639.
(39) Wiebelhaus, N. J.; Cranswick, M. A.; Klein, E. L.; Lockett, L. T.; Lichtenberger, D.
L.; Enemark, J. H. Inorg. Chem. 2011, 50, 11021.
(40) Wolpher, H.; Borgström, M.; Hammarström, L.; Bergquist, J.; Sundström, V.;
Styring, S.; Sun, L.; Åkermark, B. Inorg. Chem. Commun. 2003, 6, 989.
(41) Na, Y.; Wang, M.; Pan, J.; Zhang, P.; Åkermark, B.; Sun, L. Inorg. Chem. 2008, 47,
2805.
(42) Li, X.; Wang, M.; Zhang, S.; Pan, J.; Na, Y.; Liu, J.; Aakermark, B.; Sun, L. J Phys
Chem B 2008, 112, 8198.
(43) Te Velde, G.; Bickelhaupt, F. M.; Baerends, E. J.; Fonseca Guerra, C.; Van
Gisbergen, S. J. A.; Snijders, J. G.; Ziegler, T. J. Comput. Chem. 2001, 22, 931.
(44) Swart, M.; Ehlers, A. W.; Lammertsma, K. Mol. Phys. 2004, 102, 2467.
(45) Zhang, Y.; Wu, A.; Xu, X.; Yan, Y. Chem. Phys. Lett. 2006, 421, 383.
(46) Macias-Ruvalcaba, N. A.; Evans, D. H. J Phys Chem B 2005, 109, 14642.
(47) Felton, G. A. N.; Glass, R. S.; Lichtenberger, D. L.; Evans, D. H. Inorg. Chem.
2006, 45, 9181.
(48) Rudolph, M. J. Electroanal. Chem. 2003, 543, 23.
(49) Stoll, S.; Schweiger, A. J. Magn. Reson. 2006, 178, 42.
(50) Guerra, C. F.; Snijders, J. G.; Te Velde, G.; Baerends, E. J. Theor. Chem. Acc. 1998,
99, 391.
(51) ADF2012.01 2012
152
(52) Neese, F. WIREs Comput Mol Sci 2012, 2, 73.
(53) Frisch, M. J., et al Gaussian 09, Revision B.01, Gaussian, Inc., Wallingford CT
2009, , TY: GEN; ID: citeulike:9096580.
(54) Stoll, H.; Pavlidou, C. M. E.; Preuss, H. Theor. Chim. Acta 1978, 49, 143.
(55) Humphrey, W.; Dalke, A.; Schulten, K. J. Molec. Graphics 1996, 14, 33.
(56) van Lenthe, E.; Ehlers, A.; Baerends, E. J. Chem. Phys. 1999, 110, 8943.
(57) Zavarine, I. S.; Kubiak, C. P. J Electroanal Chem 2001, 495, 106.
(58) Hall, G. B.; Chen, J.; Mebi, C. A.; Okumura, N.; Swenson, M. T.; Ossowski, S. E.;
Zakai, U. I.; Nichol, G. S.; Lichtenberger, D. L.; Evans, D. H.; Glass, R. S.
Organometallics 2013, 32, 6605.
(59) Adams, M. W. Biochim. Biophys. Acta 1990, 1020, 115.
(60) Hatchikian, E. C.; Forget, N.; Fernandez, V. M.; Williams, R.; Cammack, R. Eur. J.
Biochem. 1992, 209, 357.
(61) Allen, F. H. Acta Crystallogr. , Sect. B: Struct. Sci. 2002, B58, 380.
(62) Bruno, I. J.; Cole, J. C.; Edgington, P. R.; Kessler, M.; Macrae, C. F.; McCabe, P.;
Pearson, J.; Taylor, R. Acta Crystallogr. , Sect. B: Struct. Sci. 2002, B58, 389.
(63) Evidence for the electronic coupling of the 2Fe2S and 4Fe4S via sulfur connecting
these moieties has been presented: Schwab, D. E.; Tard, C.; Brecht, E.; Peters, J. W.;
Pickett, C. J.; Szilagyi, R. K. Chem. Commun. 2006, 35, 3696.
(64) Recently an example has been reported in which a fullerene is attached to an iron of
a 2Fe2S core via phosphorus. Evidence for electron delocalization between the
fullerene and Fe moieties in its 1e- reduction product was presented. Liu,Y.C.; Yen,
T.; Tseng, Y.; Hu, C.; Lee, G.; Chiang, M. Inorg. Chem. 2012, 51, 5997.
(65) Adams, R. D.; Miao, S. Inorg. Chem. 2004, 43, 8414.
(66) Pierpont, C. G.; Lange, C. W. Prog. Inorg. Chem. 1994, 41, 331.
(67) Chaudhuri, P.; Wieghardt, K. Prog. Inorg. Chem. 2001, 50, 151.
(68) De Bruin, B.; Hetterscheid, D. G. H.; Koekkoek, A. J. J.; Gruetzmacher, H. Prog.
Inorg. Chem. 2007, 55, 247.
(69) Kaim, W. Inorg. Chem. 2011, 50, 9752.
153
(70) Chen, J.; Vannucci, A. K.; Mebi, C. A.; Okumura, N.; Borowski, S. C.; Swenson,
M.; Lockett, L. T.; Evans, D. H.; Glass, R. S.; Lichtenberger, D. L. Organometallics
2010, 29, 5330.
(71) Peover, M. E. J. Chem. Soc. 1962, , 4540.
(72) All pKa values taken from:Izutsu, K. Acid-Base Dissociation Constants in Dipolar
Aprotic Solvents; Blackwell Scientific Publishers: Oxford, UK, 1990.
(73) This explanation is approximate. In actuality, the ratio of acid to catalyst that is
sufficient to produce the catalytic reduction peak near -1.9 V will depend upon the
relative diffusion coefficients of acid and catalyst. , , .
(74) Pedersen, J. A., Ed.; In CRC Handbook of EPR Spectra from Quinones and Quinols;
CRC Press, Inc.: Boca Raton, Florida, 1985; .
(75) Gough, T. E. Trans. Faraday Soc. 1966, 62, 2321.
(76) Holton, D. M.; Murphy, D. J. Chem. Soc. , Faraday Trans. 1 1982, 78, 1223.
(77) Wang, W.; Nilges, M. J.; Rauchfuss, T. B.; Stein, M. J. Am. Chem. Soc. 2013, 135,
3633.
(78) Justice, A. K.; Nilges, M. J.; Rauchfuss, T. B.; Wilson, S. R.; De Gioia, L.;
Zampella, G. J. Am. Chem. Soc. 2008, 130, 5293.
(79) Schilter, D.; Rauchfuss, T. B.; Stein, M. Inorg. Chem. 2012, 51, 8931.
(80) Seidel, R. A.; Hall, G. B.; Swenson, M. T.; Nichol, G. S.; Lichtenberger, D. L.;
Evans, D. H.; Glass, R. S. J. Sulfur Chem. 2013, 34, 566.
(81) Shivakumar, M.; Pramanik, K.; Bhattacharyya, I.; Chakravorty, A. Inorg. Chem.
2000, 39, 4332.
(82) Cabeza, J. A.; Martinez-Garcia, M. A.; Riera, V.; Ardura, D.; Garcia-Granda, S.
Organometallics 1998, 17, 1471.
(83) Hall, G. B.; Kottani, R.; Felton, G. A. N.; Yamamoto, T.; Evans, D. H.; Glass, R. S.;
Lichtenberger, D. L. J. Am. Chem. Soc. 2014, , .
(84) Antonello, S.; Maran, F. Chem. Soc. Rev. 2005, 34, 418.
(85) Houmam, A. Chem. Rev. 2008, 108, 2180.
(86) Wakamiya, A.; Nishinaga, T.; Komatsu, K. J. Am. Chem. Soc. 2002, 124, 15038.
154
(87) Block, E.; Birringer, M.; DeOrazio, R.; Fabian, J.; Glass, R. S.; Guo, C.; He, C.;
Lorance, E.; Qian, Q.; Schroeder, T. B.; Shan, Z.; Thiruvazhi, M.; Wilson, G. S.;
Zhang, X. J. Am. Chem. Soc. 2000, 122, 5052.
(88) Zhu-Ohlbach, Q.; Gleiter, R.; Rominger, F.; Schmidt, H.; Reda, T. Eur. J. Org.
Chem. 1998, 11, 2409.
(89) Schroth, W.; Dunger, S.; Billig, F.; Spitzner, R.; Herzschuh, R.; Vogt, A.; Jende, T.;
Israel, G.; Barche, J.; Strohl, D.; Sieler, J. Tetrahedron 1996, 52, 12677.
(90) Dakova, B.; Carbonnelle, P.; Walcarius, A.; Lamberts, L.; Evers, M. Electrochim.
Acta 1992, 37, 725.
(91) Hennig, H.; Schumer, F.; Reinhold, J.; Kaden, H.; Oelssner, W.; Schroth, W.;
Spitzner, R.; Hartl, F. J Phys Chem A 2006, 110, 2039.
(92) Daasbjerg, K.; Jensen, H.; Benassi, R.; Taddei, F.; Antonello, S.; Gennaro, A.;
Maran, F. J. Am. Chem. Soc. 1999, 121, 1750.
(93) Antonello, S.; Benassi, R.; Gavioli, G.; Taddei, F.; Maran, F. J. Am. Chem. Soc.
2002, 124, 7529.
(94) Antonello, S.; Daasbjerg, K.; Jensen, H.; Taddei, F.; Maran, F. J. Am. Chem. Soc.
2003, 125, 14905.
(95) Meneses, A. B.; Antonello, S.; Arevalo, M. C.; Gonzalez, C. C.; Sharma, J.;
Wallette, A. N.; Workentin, M. S.; Maran, F. Chem. Eur. J. 2007, 13, 7983.
(96) Maran, F.; Wayner, D. D. M.; Workentin, M. S. Adv. Phys. Org. Chem. 2001, 36,
85.
(97) Yamaji, M.; Tojo, S.; Takehira, K.; Tobita, S.; Fujitsuka, M.; Majima, T. J. Phys.
Chem. A 2006, 110, 13487.
(98) Lawrence, C. C.; Bennati, M.; Obias, H. V.; Bar, G.; Griffin, R. G.; Stubbe, J. Proc.
Natl. Acad. Sci. U. S. A. 1999, 96, 8979.
(99) Zipse, H.; Artin, E.; Wnuk, S.; Lohman, G. J. S.; Martino, D.; Griffin, R. G.;
Kacprzak, S.; Kaupp, M.; Hoffman, B.; Bennati, M.; Stubbe, J.; Lees, N. J. Am.
Chem. Soc. 2009, 131, 200.
(100) Gross, E.; Sevier, C. S.; Vala, A.; Kaiser, C. A.; Fass, D. Nature Struct. Biol. 2002,
9, 61.
(101) Dumont, E.; Loos, P.; Assfeld, X. Chem. Phys. Lett. 2008, 458, 276.
155
(102) Sawicka, A.; Skurski, P.; Hudgins, R. R.; Simons, J. J. Phys. Chem. B 2003, 107,
13505.
(103) Dumont, E.; Loos, P.; Assfeld, X. J. Phys. Chem. B 2008, 112, 13661.
(104) Rickard, G. A.; Berges, J.; Houee-Levin, C.; Rauk, A. J. Phys. Chem. B 2008, 112,
5774.
(105) Breitzer, J. G.; Smirnov, A. I.; Szczepura, L. F.; Wilson, S. R.; Rauchfuss, T. B.
Inorg. Chem. 2001, 40, 1421.
(106) Zweig, A.; Hoffmann, A. K. J. Org. Chem. 1965, 30, 3997.
(107) Wenska, G.; Filipiak, P.; Asmus, K.; Bobrowski, K.; Koput, J.; Marciniak, B. J.
Phys. Chem. B 2008, 112, 10045.
(108) Summers, L. A. The bipyridinium herbicides; Academic Press: New York, 1980.
(109) Nada, A. A.; Hamed, H. A.; Barakat, M. H.; Mohamed, N. R.; Veziroglu, T. N. Int.
J. Hydrogen Energy 2008, 33, 3264.
(110) Ghica, M. E.; Brett, C. M. A. Anal. Chim. Acta 2005, 532, 145.
(111) Aulenta, F.; Canosa, A.; Majone, M.; Panero, S.; Reale, P.; Rossetti, S. Environ.
Sci. Technol. 2008, 42, 6185.
(112) Saha, S.; Stoddart, J. F. Chem. Soc. Rev. 2007, 36, 77.
(113) Andersson, S.; Zou, D.; Zhang, R.; Sun, S.; Aakermark, B.; Sun, L. Eur. J. Org.
Chem. 2009, 8, 1163.
(114) Benniston, A. C.; Harriman, A.; Li, P.; Rostron, J. P.; Harrington, R. W.; Clegg, W.
Chem. Eur. J. 2007, 13, 7838.
(115) Benniston, A. C.; Hagon, J.; He, X.; Yang, S.; Harrington, R. W. Org. Lett. 2012,
14, 506.
(116) Burns, J. A.; Whitesides, G. M. J. Am. Chem. Soc. 1990, 112, 6296.
(117) Castellà-Ventura, M.; Kassab, E. J. Raman Spectrosc. 1998, 29, 511.
(118) Ould-Moussa, L.; Poizat, O.; Catella-Ventura, M.; Buntinx, G.; Kassab, E. J. Phys.
Chem. 1996, 100, 2072.
(119) Braterman, P. S.; Song, J. I. J. Org. Chem. 1991, 56, 4678.
156
(120) Hong, S. H.; Kraiya, C.; Lehmann, M. W.; Evans, D. H. Anal. Chem. 2000, 72,
454.
(121) ADF2013.01 2013,
(122) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett. 1996, 77, 3865.
(123) van Lenthe, E.; Baerends, E. J.; Snijders, J. G. J. Chem. Phys. 1994, 101, 9783.
(124) Klamt, A. J. Phys. Chem. 1995, 99, 2224.
(125) Haxel, G. B.; Hedrick, J. B.; Orris, G. J. USGS Fact Sheet 2002, , 087-02.
(126) Kim, D.; Kim, M. Bioresour. Technol. 2011, 102, 8423.
(127) Fontecave, M.; Artero, V. C. R. Chim. 2011, 14, 362.
(128) Best, S. P.; Borg, S. J.; Wihite, J. M.; Razavet, M.; Pickett, C. J. Chem. Commun.
2007, 4348.
(129) Borg, S. J.; Behrsing, T.; Best, S. P.; Razavet, M.; Liu, X.; Pickett, C. J. J. Am.
Chem. Soc. 2004, 126, 16988.
(130) Greco, C.; Zampella, G.; Bertini, L.; Bruschi, M.; Fantucci, P.; De Gioia, L. Inorg.
Chem. 2007, 46, 108.
(131) Lyon, E. J.; Georgakaki, I. P.; Reibenspies, J. H.; Darensbourg, M. Y. J. Am. Chem.
Soc. 2001, 123, 3268.
(132) Mishra, A.; Pootrakulchote, N.; Wang, M.; Moon, S.; Zakeeruddin, S. M.; Graetzel,
M.; Baeuerle, P. Adv. Funct. Mater. 2011, 21, 963.
(133) Xiang, N.; Huang, X.; Feng, X.; Liu, Y.; Zhao, B.; Deng, L.; Shen, P.; Fei, J.; Tan,
S. Dyes Pigm. 2010, 88, 75.
(134) Ewbank, P. C.; Stefan, M. C.; Sauve, G.; McCullough, R. D. In Handbook of
Thiophene-Based Materials; John Wiley & Sons: 2009; pp 157-217.
(135) Ott, S.; Kritikos, M.; Åkermark, B.; Sun, L. Angew. Chem. Int. Ed. 2003, 42, 3285.
(136) Reisner, E.; Powell, D. J.; Cavazza, C.; Fontecilla-Camps, J. C.; Armstrong, F. A.
J. Am. Chem. Soc. 2009, 131, 18457.
(137) Seyferth, D.; Henderson, R. S.; Song, L.; Womack, G. B. J. Organomet. Chem.
1985, 292, 9.
157
(138) Adams, R. D.; Cotton, F. A.; Cullen, W. R.; Hunter, D. L.; Mihichuk, L. Inorg.
Chem. 1975, 14, 1395.
(139) Yu, Z.; Wang, M.; Li, P.; Dong, W.; Wang, F.; Sun, L. Dalton Trans. 2008, 18,
2400.
(140) Si, Y.; Hu, M.; Chen, C. Comp. Rend. Chim. 2008, 11, 932.
(141) Darchen, A.; Mousser, H.; Patin, H. Chem. Commun. 1988, 968.
(142) Song, L.C., J.; Yan, J.; Wang, H.L, X.; Hu, Q. Organometallics 2006, 25, 1544.
(143) Chong, D.; Georgakaki, I. P.; Mejia-Rodriguez, R.; Sanabria-Chinchilla, J.;
Soriaga, M. P.; Darensbourg, M. Y. J. Chem. Soc. Dalton Trans. 2003, 4158.
(144) Eckert, F.; Leito, I.; Kaljurand, I.; Kuett, A.; Klamt, A.; Diedenhofen, M. J.
Comput. Chem. 2009, 30, 799.
(145) Sheldrick, G. M. Acta Cryst. 2008, A64, 112.
(146) Dolomanov, O. V.; Bourhis, L. J.; Gildea, R. J.; Howard, J. A. K.; Puschmann, H.
J. Appl. Crystallogr. 2009, 42, 339.
(147) Pure Appl. Chem. 1984, 56, 461.
(148) Lichtenberger, D. L.; Copenhaver, A. S. J. Electron Spectrosc. Relat. Phenom.
1990, 50, 335.
(149) ADF2010.02 2010,
(150) Ceperley, D. M.; Alder, B. J. Phys. Rev. Lett. 1980, 45, 566.
(151) Vosko, S. H.; Wilk, L.; Nusair, M. Can. J. Phys. 1980, 58, 1200.
(152) Savéant, J. J. Am. Chem. Soc. 1987, 109, 6788.
(153) Savéant, J. Adv. Phys. Org. Chem. 2000, 35, 117.
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement