NORTHWESTERN UNIVERSITY
Singlet Exciton Fission: A Discussion of the Mechanism and What It Means For Dye Sensitized Solar Cells
A DISSERTATION
SUBMITTED TO THE GRADUATE SCHOOL
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS
for the degree
DOCTOR OF PHILOSOPHY
Field of Chemistry
By
Eric Chapman Greyson
EVANSTON, IL
December 2007 2
© Copyright by Eric Chapman Greyson 2007 All Rights Reserved 3 ABSTRACT
A Theoretical Analysis of Singlet Exciton Fission for Solar Cell Applications: Elucidating the Mechanism and Molecular Design
Eric Chapman Greyson
This thesis investigates singlet exciton fission, a physical process that converts one singlet exciton to a pair of triplet excitons. Singlet fission was discovered nearly forty years ago, but the mechanism for this process is still not well understood. Recent work has suggested that singlet fission may be capable of enhancing the performance of dye-sensitized solar cells. This thesis proposes the first complete mechanism for singlet fission and examines what type of molecules will undergo efficient singlet fission.
We begin by reviewing the literature of singlet fission and by examining individual chromophores that have been identified as promising for singlet fission. Electronic structure methods, such as density functional theory, are used to examine how the properties of individual chromophores are affected by combining them to create a coupled chromophore pair (CCP). We focus on how the energy levels and electronic coupling depend on the specific geometry of these
CCPs, and how those parameters are expected to affect singlet fission.
We propose a mechanism for singlet fission whereby the initial singlet excitation undergoes one electron transfer event to reach a charge transfer intermediate, followed by a second electron transfer to produce the triplet pair state. A ten state model is developed in order to analyze the dynamics and efficacy of singlet fission in molecular systems. We examine the dynamics of several molecules, both real and simulated, within this model in two different regimes of electron transfer. Density matrix formalism is used first, to examine how singlet 4 fission proceeds when the electron transfer steps are fast and coherent. We then use Marcus theory and classical kinetics to investigate singlet fission in the regime of slow electron transfer steps, where molecular rearrangement and vibrations play a critical role in electron transfer. We use each set of simulations to predict principles of molecular design for ideal CCPs for singlet fission.
5 ACKNOWLEDGEMENTS
This thesis would not have been written without the help and support of a great many individuals I have been fortunate to work with, live with, and know over the past five years. On an academic level, I have been blessed with many wonderful interactions with professors and coworkers. The opportunity to study this fascinating problem was provided by Professor Mark
Ratner, who has been a tremendous advisor, and a great encouragement throughout graduate school. This project, and theoretical chemistry in general, provided a daunting challenge for me, and it is because of Mark and the incredible community of researchers he has helped gather on the fourth floor of Ryan Hall that I was able to have the success I did. I also owe thanks to many other professors for their support and education over the years. Chiefly, I owe Teri Odom for providing me with the opportunity to investigate various problems with the realm of nanoscience, and my committee members, Tobin Marks, Michael Wasielewski, and Jim Ibers, all of whom provided insight and guidance along the way. In addition, I’ve enjoyed and learned from several conversations with Vladimiro Mujica, Abe Nitzan, Josef Michl, Troy Van Voorhis,
Franz Geiger, Ken Poeppelmeier, and Karl Scheidt.
I am constantly amazed at how much help I have received from several coworkers, and also at how many of my coworkers have become my best friends. In my life as a theoretical chemist, Sina Yeganeh has acted as an all-purpose resource, available for the discussion of anything theoretical, as long as I don’t start talking about real molecules. Irina Paci helped me begin this adventure in singlet fission and has remained helpful and encouraging even after she left. Chad Risko, Eric Brown, Thorsten Hansen, Christine Aikens and Tony Dutoi have all helped me gain a greater understanding of electronic structure theory and how it can be used 6 properly and improperly. I owe Josh Vura-Weis for his work implementing several codes, as
well as for several helpful discussions. I also owe thanks to several wonderful officemates who
have made it a pleasure to come in to work every day, Hyonseok Hwang, Sharon Koh, Marcel
Fallet, and Jen Roden. My previous lifetime as an experimental chemist was full of just as many
helpful and kind people, most notably Liza Babayan, Jeremy Barton, Chris Stender, Joel Henzie
and Perumal Sekar. Beyond this list, an incredible number of other students and friends have
helped me out in innumerable ways. Every person on this list has been not only a great help
academically, but a good friend.
Outside of work, I have had the pleasure to live with a parade of chemists and a pair of
non-chemists, all of whom have provided support and a feeling of home during the transient time
that is grad school. Several of these people could be mentioned elsewhere in these
acknowledgements, but I now acknowledge Randall Goldsmith, Chris Konek, Rick Kelley,
Khalid Salaita, James Cameron, Mark Witschi, Brandon Rodriguez, Julia Chamberlain, Remy
Zebrowski and Thomas Albrecht.
I also want to thank my family for their love and support. My parents and sister
provided me with wonderful guidance throughout and childhood, and continue to do so in my burgeoning adulthood. Obviously I would not be who I am, or where I am, without everything they have given me.
Finally, I’d like to thank my future family, Andrea, Diana, Barbara and David Voges.
Andrea has been a better friend and companion than I could have imagined possible throughout graduate school. I look forward to spending more time with her, and with the entire Voges clan, who have provided a wonderful second home for me. 7 Table of Contents
ABSTRACT...... 3
ACKNOWLEDGEMENTS ...... 5
Table of Contents ...... 7
List of Figures...... 10
List of Tables ...... 13
List of Tables ...... 13
Chapter 1 Introduction...... 14
1.1 Context and Motivation...... 15
1.2 Dye Sensitized Solar Cells ...... 15
1.3 Experimental Observation of Singlet Fission ...... 16
1.4 Design of Molecular Systems for Singlet Fission ...... 18
1.5 Outline of Thesis ...... 18
Chapter 2 Investigating the Effects of Coupling Chromophores on Electronic Structure: How does Coupled Chromophore Pair Geometry Control Singlet Fission?...... 21
2.1 Introduction...... 22
2.2 Methodology ...... 25
2.3 Results and Discussion ...... 28
2.3.1 DPIBF CCPs ...... 28 2.3.2 BQD CCPs...... 34 2.3.3 Tetracene CCPs...... 38 2.3.4 Hetero-CCPs ...... 42 2.3.5 Trade-Off of Transfer Integrals with Thermodynamics ...... 45 2.3.6 Comparison to Experimental Data...... 46
2.4 Summary and Conclusion ...... 48 8 Chapter 3 Development of a Mechanism for Singlet Fission and Analysis of Fission Dynamics and Yield In The Regime of Coherent Electron Transfer...... 49
3.1 Introduction...... 50
3.2 Methodology ...... 50
3.2.1 Creation of a Model System ...... 51 3.2.2 Electronic Structure Computations...... 56 3.2.3 Time Evolution of Excitation...... 59
3.3 Results and Discussion ...... 61
3.3.1 Electronic Structure ...... 61 3.3.2 Time Evolution of Model Systems ...... 65 3.3.2.1 Basic Model ...... 65 3.3.2.2 Effect of Decay Rates on Singlet Fission ...... 68 3.3.2.3 Effect of Matrix Elements on Singlet Fission...... 71 3.3.2.4 Effect of Non-degenerate State Energies on Singlet Fission...... 75 3.3.2.5 Lessons for Simulations...... 79 3.3.3 Time Evolution of Real Molecules...... 79
3.4 Summary and Conclusion ...... 81
Chapter 4 Singlet Fission Yield and Dynamics In the Regime of Incoherent Electron Transfer ...... 83
4.1 Introduction...... 84
4.2 Methodology ...... 84
4.2.1 Electronic Structure Computations...... 85 4.2.2 Time Evolution of the Excitation...... 86
4.3 Results and Discussion ...... 89
4.3.1 Electronic Structure Computations...... 89 4.3.2 Time Evolution of Model Systems ...... 93 4.3.2.1 Basic System...... 93 4.3.2.2 Effect of Changing Matrix Elements on Singlet Fission ...... 96 4.3.2.3 Non-Isoenergetic State Schemes...... 97 9 4.3.2.4 Conclusions from Model Simulations ...... 103 4.3.3 Time Evolution of Real Molecules...... 103
4.4 Design Principles for Singlet Fission in DSSCs...... 105
4.5 Limits of the Model ...... 107
4.6 Perspective for Experimentalists ...... 108
4.7 Summary and Conclusion ...... 112
References...... 113
Appendix A Fortran Codes...... 120
A-1 Fortran (f90) Code To Compute Orbital Overlap ...... 121
A-2 Fotran (f90) Code for Coherent Dynamics...... 125
A-3 Fortran (f90) Code for Marcus Theory Dynamics...... 139 10 List of Figures
Figure 1 HOMO, LUMO and chemical structure (Left to Right) of DPIBF, BQD and tetracene (Top to Bottom)...... 24
Figure 2 The electron transfer integrals, t, for the HOMO and LUMO are half the splitting of the CCP orbitals in a homo-CCP...... 26
Figure 3 Chemical Structures of DPIBF CCPs...... 31
Figure 4 Frontier molecular orbitals (Left to Right, HOMO-1, HOMO, LUMO, LUMO+1) for both strongly (D20, Top) and weakly (D2, Bottom) coupled DPIBF CCPs look like a weakly perturbed sum and difference of IC orbitals. The HOMO-1and LUMO have bonding interactions where the ICs join, while the HOMO and LUMO+1 have antibonding interactions at the coupling...... 33
Figure 5 Chemical Structures of BQD CCPs...... 35
Figure 6 The frontier molecular orbitals of B9 (Left to Right, HOMO-1, HOMO, LUMO, LUMO+1) are linear combinations of the isolated BQD orbitals. Because BQD has more electron density on the bridging carbons in its HOMO than its LUMO, th is much greater than tl...... 37
Figure 7 Chemical Structures of tetracene CCPs...... 39
Figure 8 The Frontier orbitals of T4 are linear combinations of the frontier orbitals of two isolated chromophores. (Left to Right, HOMO-1, HOMO, LUMO, LUMO+1)...... 41
Figure 9 Chemical structures of Hetero-CCPs...... 43
Figure 10 Frontier orbitals for the hetero-CCP DT2 are localized, in contrast to the delocalized orbitals seen in homo-CCPs. (Left to Right, Top row then Bottom HOMO-1, HOMO, LUMO, LUMO+1)...... 44
Figure 11 The S 1 excitation for a hetero-CCP (Center) is less energetic than either IC’s S 1 excitation (Left and Right), making fission more endothermic in hetero-CCPs than in homo-CCPs with comparable coupling...... 45
Figure 12 The free energy of singlet fission (Eqn 2.5, Table 1) becomes more endothermic nearly linearly as the coupling between two DPIBF chromophores is increased ...... 46
Figure 13 There are 70 possible electron configurations in a 4 electron 4 orbital basis comprised of a pair of chromophore HOMOs and LUMOs. 36 states have two neutral chromophores, 32 are a single electron transfer charge transfer pair, and 2 states put all 11 four electrons on one chromophore. Highlighted states are those that are most energetically accessible...... 52
Figure 14 Ten state system for singlet fission with all allowed electron transfers shown. States have been renumbered...... 54
Figure 15 Structures of three promising CCPs. We examine these three CCPs, as well as some variants that have amino (a), or nitro (n) ‘R’ groups...... 62
Figure 16 Fission yield is ~11% if all states are isoenergetic, all matrix elements are equal, and all decay pathways are equal. The population oscillates between different states very quickly, with fission yield based on average state populations and the ratio of the rate- limiting singlet and triplet decay processes. The inset shows the first 0.1 ps in detail. S1, CT, and TT labels refer to the sum of all population in a S1, CT or TT state. Fission and Fluorescence are the percent of the population that has left by either triplet pair or singlet decay routes, and Percent is the percent of population that has left the system through the fission route as opposed to the singlet decay route...... 65
Figure 17 (Left) Doubled triplet-triplet decay rates lead to increased fission yield, while (Right) doubled singlet decay rates decrease fission yield...... 67
Figure 18 (Left) Even with no possible singlet decay route, only 25% of the excitation undergoes fission because of destructive interference if all states have equal energy and all matrix elements are equal, (Center) If states 1 and 4 are artificially removed from the system, the fission yield is 50% under these conditions, (Right) With only one S 1, one CT, and one TT state, the fission yield is 100%...... 69
Figure 19 Doubling all of the matrix elements in an isoenergetic system with slow decay rates does not affect fission yield, although it does increase the frequency of population oscillation ...... 72
Figure 20 If the symmetry of the system is broken by raising some matrix elements and not others, the fission yield increases. (Left) If TL is doubled, the fission yield increases to 19%, (Center) If TD1 is doubled the fission yield increases to 23%, (Right) If both are doubled, fission yield increases to 27%...... 73
Figure 21 A moderate CT barrier does not affect fission yield if the S 1 and TT states are isoenergetic, although the period of oscillation increases with larger barriers (Left) 0 .081 eV barrier, and (Right) 0.27 eV...... 75
Figure 22 If the S 1 and TT states are not isoenergetic, the fission yield is decreased. If there is a G, then the presence of a moderate CT barrier does hurt fission yield. (Left) G =0.027 eV. (Center) G =0.027 eV and CT=0.081 eV (Right) G =0.027 and CT=0.27 eV ...... 77
Figure 23 Hetero-CCPs show promise because the different energy levels of the left and right localized S 1 and CT states lead to higher fission yields. (Left) States 1,4 are raised 12 0.027 eV above all others. (Center) States 5,8 are raised 0.027 eV. (Right) States 1,4,5,8 are all raised 0.027 eV...... 78
Figure 24 If all states are isoenergetic, at equilibrium each of the ten states is evenly populated. This leads to 33% fission yield, if the singlet and triplet decay rates are equal, because twice as much population resides in the four S 1 states as in the two TT states. Equilibrium is reached very quickly when all states are isoenergetic...... 94
Figure 25 Early time profiles show that larger electronic coupling matrix elements lead to quicker equilibration between states. (Left) All coupling elements .027 eV (Center) TH and TL are .054 eV so S 1 and CT equilibrate faster (right) All coupling elements are .054 eV, so all states equilibrate faster ...... 97
Figure 26 When not all states are isoenergetic their relative populations change. (Left) 0.054 eV CT state with no decay leads to each CT state having less population than the S 1 and TT states, (Center) With decay, the fission yield is only slightly depressed for this small barrier value, (Right) but a CT barrier of 0.11 eV significantly impedes fission...... 99
Figure 27 (Left) If the TT states are higher (0.027 eV) in energy than the S 1 states, less population goes to the TT states and the fission yield decreases (Center) while a low energy (-0.027 eV) TT state leads to more TT population, and higher fission rate. (Right) Adding a CT barrier (0.11 eV) does not affect the equilibrium population ratio of S 1 and TT states, but it causes more population decay out of the S 1 states before equilibrium is reached, lowering fission yield ...... 100
Figure 28 (Left) In a model hetero-CCP with the S 1 states at 0 and 0.081 eV, the CT states at - 0.027 and 0.135 eV and the TT states at -0.027 eV the fission yield is very high. (Right) If the low energy localized singlet cannot inject into the electrode, the fission yield jumps to nearly 100%...... 102
Figure 29 (Left) anPOLY has a two CT states lower in energy than its S 1 states, and has a TT state that is even lower in energy, providing excellent fission yield. (Center) anPENT also has two CT states that are lower in energy than its S 1 states, but its TT states are also above these low CT states. If there is no decay route directly from the CT states, this will eventually provide a high fission yield, (Right) but even a very slow CT decay route makes this energy scheme poor, and results in ...... 105
Figure 30 Flowchart for designing promising molecules for singlet fission. The important molecular properties are displayed at right, with possible means of achieveing those properties at left...... 109 13 List of Tables
Table 1 Electronic Matrix Elements and Free Energies for DPIBF CCPs ...... 29
Table 2 Electronic Matrix Elements and Free Energies for CCPs of BQD...... 34
Table 3 Electronic Matrix Elements and Free Energies for tetracene CCPs ...... 40
Table 4 Electronic Matrix Elements and Free Energies for Hetero-CCPs ...... 43
Table 5 Electronic Matrix Elements for Several Bare and Functionalized CCPs. The prefixes aa, an, and nn are diamino, aminonitro, and dinitro functionalized variants...... 63
Table 6 Energies of Electronic States for several CCPs (in eV)...... 64
Table 7 Summary of several simulations with various parameters for coherent dynamics. Each row is a different simulation, and the figure in which more details can be found is listed. The electronic coupling matrix elements are listed first, then the decay rate constants, followed by the energies of the states (E14 is states 1 and 4), and finally the fission percent is listed. More detail on every simulation is in the text...... 68
Table 8 Fission Yield of Real CCPs with State Energies Provided (in eV) ...... 80
Table 9 Energy of 16 ICs at every Relevant Electronic State and Geometry. Molecules are labeled by IC type (PO=polyene, DP=DPIBF, PE=pentacene) and functional group (a=amino, h=hydroxy, n=nitro). All values are in eV...... 91
Table 10 Energies for 16 CCPs for each state and geometry. The CDFT correction energy for CT states has been applied. All energies are given in eV relative to the lowest S 1S0 state. Each state is listed as a State@Geometry. The CCP scaffolds are abbreviated PO=polyene, DP=DPIBF, PE=pentacene, with variants aa=diamino, hh=dihydroxy, nn=dinitro, an=aminonitro...... 92
Table 11 Summary of several simulations with various parameters for incoherent dynamics. Each row is a different simulation, and the figure in which more details can be found is listed. The electronic coupling matrix elements are listed first, then the decay rate constants, followed by the energies of the states (E14 is states 1 and 4), and finally the fission percent is listed. More detail on every simulation is in the text ...... 95
14
Chapter 1
Introduction
15 1.1 Context and Motivation
One of the greatest challenges of the current century is sustainability. One aspect of sustainability is global energy use, which has rapidly grown in recent years, and promises to continue its rapid ascent. Various scenarios predict a global energy use of 26.4 to 32.9 terawatts
(TW) by 2050, up from only 12.7 TW in 1998.1 While a majority of energy currently comes from fossil fuels, there are several reasons to look for alternate energy sources, ranging from possibly dire environmental consequences from climate change and pollution, to the economic cost of supplying and distributing a rapidly increasing amount of fuel from a global supply that is, if anything, becoming more difficult to extract. 2, 3 The amount of energy that strikes the Earth
from sunlight in one hour is more than the amount of energy used by the global population in a
whole year. The sun clearly provides enough energy, if we are only able to harness it efficiently
and economically. This abundant supply makes solar energy one of the most promising
alternative energy sources. 4-6
1.2 Dye Sensitized Solar Cells
One of the most promising solar cells being studied is the dye sensitized solar cell
(DSSC), also known as the Grätzel cell. DSSCs are promising because they combine low cost
fabrication with potentially high energetic yield. 7, 8 A DSSC is composed of a nanoscale porous metal oxide semiconductor, coated with an organic dye, and embedded in a redox electrolyte.
The organic dye is excited by photon of light. These excited state dye molecules then each inject
(ideally) one electron into the conduction band of the metal oxide. The electrons quickly thermalize to the bottom of the conduction band as they migrate through the metal oxide to an 16 electrode. The electrons continue through the external circuit, providing power, and emerge
from another electrode at lower potential. These lower potential electrons are then recombined
with the dye molecule to reform the ground state of the molecule by a redox shuttle. The highest
reported yield for a Grätzel cell is ~11%,8-10 and this value has not increased for several years despite multiple innovative attempts. 9, 11-13 These cells suffer from many avenues of loss, but one of the most significant is the loss that comes from using a dye sensitizer that can, at best, yield one electron of current per photon absorbed. Based on the solar spectrum and a single absorber such as this, the thermodynamic limit for solar cell efficiency is 32%. 14 One possible way to improve the energetic yield of a DSSC is to use an absorber that is capable of injecting multiple electrons per absorbed photon. One process that would enable this is multiple exciton generation, which has been observed in various inorganic nanocrystals. 15-18 A molecular analog of this process is singlet fission, a physical process that converts one singlet excitation to two triplet excitations. It has been pointed out that a DSSC using a dye capable of efficient singlet fission could provide a 46% energetic yield by injecting two triplets instead of one singlet excitation for all green, blue and ultraviolet light. 14
1.3 Experimental Observation of Singlet Fission
Singlet exciton fission was first discovered in solid state polyacenes in 1965.19 It was determined that the main nonradiative decay in tetracene was triplet formation via singlet fission.
Subsequent studies examined the rate constant for singlet fission in a few solid state polyacene systems 20-29 and explored the unique magnetic field dependence of fission. 30-36 This work also led to the discovery that fission primarily occurs when the energy of the first excited singlet is 17 37-39 greater than twice the energy of the first excited triplet ( E(S 1) ≥ 2 E(T 1)). In the last
dozen years, high triplet yields attributed to singlet fission have been observed in several
crystalline systems, such as benzophenone 40 and p-sexiphenyl. 41 Singlet fission has also been
suggested in a number of polymeric systems such as polydiacetylene, poly( p-
phenylenevinylene), and a ladder-type poly( p-phenylene). 42-49 Recent studies have suggested
singlet fission in molecular systems consisting of weakly coupled chromophores such as
carotenoids, tetracene and diphenylisobenzofurans (DPIBF). 50-54
One impediment to these studies has been the difficulty in confirming whether singlet
fission occurred. Delayed fluorescence is sometimes used as evidence that singlet fission has
occurred and then reversed course, but other physical processes can also lead to delayed
fluorescence. High triplet yield is also a possible indicator for singlet fission, but unless the
triplet yield is greater than 100%, simple observation of triplet population does not rule out other
mechanisms such as intersystem crossing. As mentioned earlier, singlet fission has a unique
magnetic field angle dependence in single crystal polyacenes owing to a level crossing resonance
between pairs of triplet excitons. No group has conducted a molecular analog to this experiment
yet, which could be imagined by using polarized light to selectively excite molecules in certain
orientations, or by poling molecules to align them, for example in a liquid crystal. 55 It is also possible that an electron paramagnetic resonance experiment could definitively identify triplet fission based on the relative populations of the various triplet sublevels. 56 As of now, other than reports in single crystal polyacenes, there have been no indisputable experimental reports of singlet fission.
18 1.4 Design of Molecular Systems for Singlet Fission
The studies of polyacenes revealed that singlet fission was observable when the energy of a singlet excitation was greater than or equal to the energy of two triplet excitations. One recent paper considered what sort of molecular motif would lead to promising chromophores for singlet fission in dye sensitized solar cells. 57 The optimal singlet energy for fission with solar radiation
and common DSSC elements would be ~ 2.2 eV, with a triplet roughly half that energy. It was
also determined that the second excited triplet energy level would ideally be above the first
singlet to prevent intersystem crossing, as many chromophores with a low energy first triplet also
have other low lying triplet levels. In a simple Hückel model, the S 1 and T 1 levels are
isoenergetic. Using a simple self consistent field approximation, the T1 energy level drops below
the S 1 energy level by twice the exchange integral. It was concluded that by maximizing this
exchange integral, by creating molecules with similar electron densities across the molecules in
both the HOMO and LUMO, one is able to generate several chromophores with low triplet
energy levels. Three molecules were chosen as a result of this work as being most promising for
singlet fission: pentacene, benzoquinodimethanes (BQD), and diphenylisobenzofuran (DPIBF). 57
1.5 Outline of Thesis
The objective of this thesis is to build up a theoretical understanding of singlet fission, so
that we can work more efficiently to design molecules and systems that are capable of efficient
fission yield. We approach the problems using the tools of theoretical chemistry and aim to
provide physical insight to the process as well as to guide future work in the field. 19 In Chapter 2 we examine the effects of coupling two isolated chromophores (ICs) to form a coupled chromophore pair (CCP). We outline a framework for understanding the relation between strong electronic coupling between chromophores, which often leads to efficient electron transfer, and reaction driving force, which plays a key role in determining whether fission will occur. Electronic structure techniques are used to scan through a variety of CCP geometries for three different ICs and to pick out the most promising CCPs for molecular singlet fission in solar cells
In Chapter 3 we consider singlet fission in a weakly coupled CCP as a four electron four orbital system, using the HOMOs and LUMOs of the ICs as our orbitals, and we derive a mechanism for singlet fission after considering all seventy possible electron configurations in this basis. The derived mechanism involves two subsequent electron transfer events, going from a localized singlet exciton to a charge transfer intermediate and ending in a pair of localized triplets. We assume the electron transfer steps are fast, and coherent in this chapter, and examine how various molecular properties affect singlet fission yield in this regime. Electronic structure methods are used to compute properties of real molecules, and the density matrix formalism is used to examine the time evolution of the initial singlet excitation. We use these simulations to derive design rules for singlet fission in a system governed by coherent electron transfer.
In Chapter 4, we re-examine the two step electron transfer mechanism in the context of
Marcus theory, where the electron transfer steps are incoherent, and their rates are mediated by fluctuations of the molecules and of their environment. We calculate energy levels and electronic coupling matrix elements for several promising molecules, and examine the time evolution of the initial excitation. We consider how molecular properties as well as the 20 properties of the environment affect singlet fission yield and derive guidelines for an optimal molecular system for singlet fission within this regime.
In summary, this work begins by considering individual chromophores, the most basic building blocks for a singlet fission enhanced dye-sensitized solar cell. We examine how the individual chromophores interact, and what control we can have over their molecular interactions. A complete mechanism for singlet fission is derived for weakly coupled chromophore pairs and we examine how the initial singlet excitation evolves and decays in both the coherent and the incoherent regimes of electron transfer. These studies are used to propose target molecules to realize high singlet fission yield in molecular systems and to move towards improved solar cell efficiency by means of singlet fission.
21
Chapter 2
Investigating the Effects of Coupling Chromophores on Electronic
Structure: How does Coupled Chromophore Pair Geometry Control
Singlet Fission?
22 2.1 Introduction
Recent experiments have examined the effects of chemically connecting two promising singlet fission chromophores in order to encourage energy transfer, and hence singlet fission.
The highest reported triplet yields in these coupled chromophore pairs (CCPs) are under 10%, as opposed to an ideal of 200% triplet yield. 50-53, 58 The exact nature of the connection, and therefore interaction, between coupled chromophores should control the efficiency of singlet fission. A CCP can be viewed as a perturbation of the electronic structure of two isolated chromophores (ICs) to a combined system, much like excimer formation in solution. Unlike excimer formation, one can control the intermolecular interaction by controlling the bonding between the two chromophores. The exact arrangement of the chromophores in space is critical: while single crystal tetracene has a very high triplet yield from singlet fission (room temperature fluorescence yield is less than 0.2%), solutions of molecular tetracene show no triplets, and the three known CCPs of tetracene to have been studied have delayed fluorescence yields from 0 to
3%, which could indicate weak fission.53, 54 Similarly, three CCPs of DPIBF have been studied, and have triplet yields that vary from 0% to 9% depending on the CCP geometry. 58
The specific chemical connection between two monomers will have two major effects.
First, it will affect the overall energetic balance of singlet fission, which is critical because fission yields will be low or zero if the process is significantly endothermic. One would expect that CCP formation would lead to some delocalization of the initial singlet exciton, and thus a stabilization of the initial singlet exciton state. The final pair of triplets state is not stabilized, however, as the two triplets are each localized on separate halves of the system. Stronger electronic communication between the two halves would be expected to result in greater 23 delocalization, hence more stabilization of the initial excited singlet, and a less energetically favorable singlet fission process. Because it is difficult to find ICs where singlet fission is significantly exothermic, it is important to monitor this property when creating a CCP. Second, the degree of electronic communication between the two halves is determined by the connection.
The figure of merit for electronic coupling is the electron transfer integral (or electronic coupling matrix element), t, and is an important factor in energy and electron transfer, which often (for small t) have rates proportional to | t|2.59-61 It is clear that the ideal CCP will have to balance boosting the kinetic driving force with harming the energetic balance of fission.
In this chapter we design and examine a series of CCPs based on tetracene, BQD and
DPIBF.(Figure 1) We investigate how various modes of chemical connectivity affect the electronic structure of the system, focusing on important parameters for singlet fission. In particular, we examine the electron transfer integral (i.e. electronic coupling), a key parameter in energy and electron transfer. We also study the free energy of singlet fission - whether singlet fission is endothermic or exothermic. Finally, we use these computations to predict specific
CCPs that are promising for singlet fission in solar cells.
24
O
N N
N N
Figure 1 HOMO, LUMO and chemical structure (Left to Right) of DPIBF, BQD and tetracene (Top to Bottom)
The three ICs we consider are large aromatic systems, a structural motif which is known to lead to low triplet energies because of large exchange stabilizations. 57 Each of the three chromophores was coupled with a partner in multiple ways to see how the electron transfer integral and free energy of singlet fission would vary based upon CCP geometry. We use the frontier molecular orbitals of the ICs to guide our study, as they provide a simple tool for predicting how strongly perturbed a CCP will be from two non-interacting chromophores. The perturbation is roughly proportional to the amount of overlap between the frontier orbitals of the two monomers. In some cases, the IC structure was altered slightly to create a larger variety of
CCP structures and parameter values.
25
2.2 Methodology
We focus on two important parameters for singlet fission, the electron transfer integral, t, and the thermodynamic balance of singlet fission, Gf, which is the energy of the final state
minus the energy of the intial state. We derive both parameters from electronic structure
computations. Electronic structure calculations for both ICs and CCPs are done at the Density
Functional Theory (DFT) level with the B3LYP functional, using the 6-31G** basis set. Time
dependent DFT (TDDFT) calculations were used to evaluate the excited singlet states for all
systems. Hartree-Fock theory and configuration interaction (HF-CIS) were used to confirm DFT
and TDDFT results. Both the ground and first excited singlet were evaluated at the optimized
geometry of the ground state singlet, while the first excited triplet state was evaluated at its own
minimum energy geometry. All calculations were performed with either QChem 3.0 62 or
NWChem 5.0.63, 64
The electron transfer integrals were approximated by invoking Koopmans’ theorem and
assuming that only the two monomer LUMOs (HOMOs) mix to form the CCP LUMO and
LUMO+1 (HOMO and HOMO-1).(Figure 2) 65 We believe this is a safe approximation because
the energetic gaps between the HOMO-1 and HOMO and LUMO and LUMO+1 are more than 1
eV for all three chromophores.
26
LUMO 1 LUMO 2 2 tl ENERGY
HOMO 1 HOMO 2 2 th
Figure 2 The electron transfer integrals, t, for the HOMO and LUMO are half the splitting of the CCP orbitals in a homo-CCP
Specifically, the 2x2 Hamiltonian:
E1 t H = (2.1) t E 2
where E1 and E2 are the energies of the HOMOs (LUMOs) for the two uncoupled chromophores, gives rise to eigenstates which are the HOMO and HOMO-1 (LUMO and
LUMO+1) of the CCP. If we take the orbital overlap to be very small, the electron transfer integral, t, can be approximated as:
2t= 4'( t2 − G ) 2 (2.2) 27
where G is the difference in the HOMO (LUMO) energies of the two ICs (E 1 and E 2)
that are mixing, and 2 t ' is the difference in the energy of the HOMO and HOMO-1 (LUMO,
LUMO+1) of the CCP. When both chromophores are the same, E1 = E2, the equation simplifies
so that t is the amount the CCP HOMO and HOMO-1 are above and below the chromophore
HOMO.
E+/ − = E ± t (2.3)
To evaluate the energetic balance of singlet fission, Gf, we assume the initial state is the lowest energy singlet exciton of the CCP. The CCP LUMO+1 orbital could be occupied after the molecule absorbs light, but the excess energy will quickly thermalize in most cases. The initial state energy is thus E , the CCP excited state as calculated by TDDFT, minus E , the S1 S0
CCP ground state as computed by regular DFT. There is no good way to compute the energy of
the pair of triplets state for the CCP. Because of this, we approximate the energy of the pair of
triplets as being twice the triplet-ground state energy gap for an IC, as both of these energies can
be solved with DFT. This leads to the equation: