Optical and magnetic resonance studies of organic materials used in photovoltaic applications

In this work we focused on the electronic processes in active materials used in organic photovoltaics. Films of several electron donors, acceptors, and their blends were investigated by means of steady state optical and magnetic resonance probes. The efficiency of organic photovoltaics depends on film morphology, charge mobility and light absorption. Therefore we studied common donor materials with very different morphology: RR P3HT (regioregular poly(3-hexylthiophene)) and RRa P3HT (regio-random poly(3-hexylthiophene)). The charge transport is affected by regioregularity and molecular weight. Consequently, we examined RR P3HT polymers with various molecular weights. We learned that a polaron band at low photon energy only appears in the photoinduced absorption spectrum of low molecular weight RR P3HT. We studied two main approaches for improving the efficiency of organic photovoltaics: modifying the lowest (highest) unoccupied (occupied) molecular orbital, LUMO (HOMO) of the donor (acceptor) materials; as well as synthesizing polymer donors with low optical gap. TAES-V is a low-band gap polymer composed of three co-polymers having the structure of „donor-acceptor-donor?. Its record power conversion efficiency (~7%) when blended with PC70BM is partially due to the significantly red-shifted absorption. Our results show that an intrachain charge transfer exciton (CTE) is long-lived in this polymer and that it persists in the blend with PC70BM. In addition we studied three fullerene derivatives. The LUMO of a fullerene derivative can be changed by the addition of functional side groups to the fullerene cage that improves the organic solar cells performance. The addition can also lead to hindering of aggregation in the films, which consequently decreases the charge transport in solar cells. In the study of polymer/fullerene blends we mixed RR P3HT with three different fullerene derivatives. We conclude that higher power conversion efficiency of a blend is mainly due to the higher LUMO level and improved open circuit voltage. We also compared DOO-PPV (H-polymer) with DOO-PPV enriched with deuterium (D-polymer). We show that hyperfine interaction is weaker in the D-polymer and that the spin relaxation rate is four times smaller than in H-polymer. Consequently, the longer spin diffusion length makes the D-polymer better suited for higher performing organic spin-valves.

OPTICAL AND MAGNETIC RESONANCE STUDIES OF ORGANIC MATERIALS USED IN PHOTOVOLTAIC APPLICATIONS by Golda Hukic-Markosian A dissertation submitted to the faculty of The University of Utah in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Physics Department of Physics and Astronomy The University of Utah December 2011 Copyright  Golda Hukic-Markosian 2011 All Rights Reserved Th e Uni v e r s i t y o f Ut a h Gr a dua t e S cho o l STATEMENT OF DISSERTATION APPROVAL The dissertation of Golda Hukic-Markosian has been approved by the following supervisory committee members: Zeev Valy Vardeny , Chair 06/28/2011 Date Approved Mikhail Raikh , Member 06/28/2011 Date Approved Christoph Boehme , Member 06/28/2011 Date Approved Calrton DeTar , Member 07/21/2011 Date Approved Michael Morse , Member 06/28/2011 Date Approved and by David Kieda , Chair of the Department of Physics and Astronomy and by Charles A. Wight, Dean of The Graduate School. ABSTRACT In this work we focused on the electronic processes in active materials used in organic photovoltaics. Films of several electron donors, acceptors, and their blends were investigated by means of steady state optical and magnetic resonance probes. The efficiency of organic photovoltaics depends on film morphology, charge mobility and light absorption. Therefore we studied common donor materials with very different morphology: RR P3HT (regioregular poly(3-hexylthiophene)) and RRa P3HT (regio-random poly(3-hexylthiophene)). The charge transport is affected by regioregularity and molecular weight. Consequently, we examined RR P3HT polymers with various molecular weights. We learned that a polaron band at low photon energy only appears in the photoinduced absorption spectrum of low molecular weight RR P3HT. We studied two main approaches for improving the efficiency of organic photovoltaics: modifying the lowest (highest) unoccupied (occupied) molecular orbital, LUMO (HOMO) of the donor (acceptor) materials; as well as synthesizing polymer donors with low optical gap. TAES-V is a low-band gap polymer composed of three co-polymers having the structure of „donor-acceptor-donor‟. Its record power conversion efficiency (~7%) when blended with PC70BM is partially due to the significantly red-shifted absorption. Our iv results show that an intrachain charge transfer exciton (CTE) is long-lived in this polymer and that it persists in the blend with PC70BM. In addition we studied three fullerene derivatives. The LUMO of a fullerene derivative can be changed by the addition of functional side groups to the fullerene cage that improves the organic solar cells performance. The addition can also lead to hindering of aggregation in the films, which consequently decreases the charge transport in solar cells. In the study of polymer/fullerene blends we mixed RR P3HT with three different fullerene derivatives. We conclude that higher power conversion efficiency of a blend is mainly due to the higher LUMO level and improved open circuit voltage. We also compared DOO-PPV (H-polymer) with DOO-PPV enriched with deuterium (D-polymer). We show that hyperfine interaction is weaker in the D-polymer and that the spin relaxation rate is four times smaller than in H-polymer. Consequently, the longer spin diffusion length makes the D-polymer better suited for higher performing organic spin-valves. To my husband Richard, my family and my dear friends who supported me in this amazing journey. CONTENTS ABSTRACT ..................................................................................................... iv ACKNOWLEDGMENTS ............................................................................... ix 1. INTRODUCTION ........................................................................................ 1 1.1 П-conjugated polymers ............................................................................................. 2 1.1.1. Model for stable excitations in -conjugated polymers .............................. 3 1.2 Major photoexcitations ............................................................................................. 5 1.2.1. Excitons ....................................................................................................... 5 1.5.1. Polaron excitations ...................................................................................... 8 1.5.2. Polaron pairs ................................................................................................ 8 1.5.3. Bipolaron excitations................................................................................. 10 1.3 Organic solar cells................................................................................................... 11 1.3.1. Single layer OPV device ........................................................................... 12 1.3.1. Bilayer OPV device ................................................................................... 12 1.3.2. Bulk heterojunction OPV device............................................................... 15 1.3.3. Power conversion efficiency ..................................................................... 15 1.4 Organic light emitting diodes.................................................................................. 17 1.4.1. Magnetic field effect ................................................................................. 20 1.5 Organic spintronics ................................................................................................. 21 1.5.1. Spin-orbit coupling .................................................................................... 21 1.5.2. Hyperfine interaction................................................................................. 22 2. EXPERIMENTAL TECHNIQUES............................................................ 24 2.1. Absorption and emission......................................................................................... 24 2.2. Continuous wave (CW) spectroscopy ..................................................................... 25 2.2.1. Photoinduced absorption ........................................................................... 25 2.2.2. Doping induced absorption ....................................................................... 27 2.3. Modulation spectroscopy ........................................................................................ 28 2.3.1. Mono- and bimolecular recombination ..................................................... 29 2.3.2. Modulation frequency response ................................................................ 31 2.3.3. Dispersive kinetics .................................................................................... 33 2.3.4. FTIR photomodulation spectroscopy ........................................................ 34 2.4. Electroabsorption .................................................................................................... 37 2.5. Photoluminescence quantum efficiency ................................................................. 41 2.6. Electron spin resonance spectroscopy .................................................................... 42 2.6.1. Introduction to theory ................................................................................ 42 2.6.1.1. Spin and orbital moments ................................................................ 42 2.6.1.2. Spin Hamiltonian ............................................................................ 45 vii 2.6.1.3. Information obtained from ESR ...................................................... 49 2.6.2. ESR kinetic studies.................................................................................... 49 2.6.2.1. Bloch equations ............................................................................... 49 2.6.2.2. CW ESR .......................................................................................... 51 2.6.2.3. Line shape ....................................................................................... 56 2.6.3. Experimental set up for ESR ..................................................................... 57 2.6.3.1. The microwave bridge .................................................................... 59 2.6.3.2. The ESR cavity ............................................................................... 60 2.6.3.3. Signal channel ................................................................................. 61 2.6.3.4. Magnetic field controller................................................................. 62 2.6.4. Light induced ESR (LESR) ....................................................................... 63 2.7. Optically detected magnetic resonance (ODMR) spectroscopy ............................. 63 2.7.1. PADMR ..................................................................................................... 64 2.7.2. PLDMR ..................................................................................................... 68 2.7.3. Experimental set-up................................................................................... 69 2.8. Polaron pair mixing under magnetic resonance ............................................ 72 3. STUDY OF PHOTOEXCITATIONS IN DONOR MATERIALS ........... 74 3.1. Poly(3-hexylthiophene)........................................................................................... 74 3.1.1. Introduction ............................................................................................... 74 3.1.1. Study of absorption and emission spectra ................................................. 76 3.1.2. Studies of photoinduced and doping induced absorption spectra ............. 77 3.1.3. PADMR spectroscopy ............................................................................... 82 3.1.4. Conclusion ................................................................................................. 85 3.2. TAES-V polymer .................................................................................................... 85 3.2.1. Introduction ............................................................................................... 85 3.2.2. Optical study of TEAS-V copolymer ........................................................ 86 3.2.3. Magnetic resonance study of TAES-V copolymer ................................... 95 3.2.4. Conclusion ............................................................................................... 103 4. ACCEPTOR MATERIALS IN ORGANIC PHOTOVOLTAICS ........... 104 4.1. Introduction ........................................................................................................... 104 4.2. PCBM. .................................................................................................................. 105 4.2.1. Introduction ............................................................................................. 105 4.2.2. Sample preparation .................................................................................. 105 4.2.3. Optical studies of PCBM film and isolated molecules ........................... 106 4.2.4. Magnetic resonance studies ..................................................................... 113 4.2.5. Conclusion ............................................................................................... 115 4.3. Bis-PCBM ............................................................................................................. 115 4.3.1. Introduction ............................................................................................. 115 4.3.2. Optical and magnetic resonance studies of bis-[60]PCBM .................... 117 4.3.3. Conclusion ............................................................................................... 122 4.4. Indene-C60 bisadduct ........................................................................................... 123 4.4.1. Introduction ............................................................................................. 123 4.4.3. Optical and magnetic resonance studies.................................................. 125 4.4.3. Conclusion ............................................................................................... 131 viii 4.5. General Conclusion ............................................................................................... 132 5. POLYMER / FULLERENE BLENDS ..................................................... 133 5.1. Introduction ........................................................................................................... 133 5.2. P3HT / PCBM blend ............................................................................................. 134 5.2.1. Introduction ............................................................................................. 134 5.2.2. Optical studies ......................................................................................... 135 5.2.3. Magnetic resonance studies ..................................................................... 140 5.2.4. Conclusion ............................................................................................... 151 5.3. RR P3HT / bis-[60]PCBM blend .......................................................................... 153 5.3.1. Introduction ............................................................................................. 153 5.3.2. Optical studies ......................................................................................... 153 5.3.1. Magnetic resonance studies ..................................................................... 159 5.3.2. Conclusion ............................................................................................... 163 5.4. RR P3HT / ICBA blend ........................................................................................ 164 5.4.1. Introduction ............................................................................................. 164 5.4.2. Optical studies ......................................................................................... 164 5.4.3. Magnetic resonance studies ..................................................................... 167 5.4.4. Conclusion ............................................................................................... 174 5.5. TAES-V/[C70]PCBM ............................................................................................ 175 5.5.1. Introduction ............................................................................................. 175 5.5.2. Optical studies ......................................................................................... 176 5.5.3. Magnetic resonance studies ..................................................................... 180 5.5.4. Conclusion ............................................................................................... 188 5.6. General conclusion................................................................................................ 189 6. ISOTOPE EFFECT IN SPIN RESPONSE OF Π-CONJUGATED POLYMER FILMS ................................................................................... 191 6.1. Introduction ........................................................................................................... 191 6.2. Polaron spin ½ resonance line broadening due to hyperfine splitting and inhomogeneity .............................................................................................................. 194 6.3. ODMR................................................................................................................... 195 6.4. Conclusion ............................................................................................................ 200 CONCLUSION ............................................................................................. 201 APPENDIX: GLOSSARY OF ABBREVIATIONS .................................... 206 REFERENCES ............................................................................................. 208 ACKNOWLEDGMENTS I would like to express my deepest gratitude to my advisor Prof. Valy Vardeny. This work wouldn‟t be possible without his constant guidance, encouragement, and care. I was so fortunate to experience and learn from his passion for understanding physics and finding scientific truth even when it differs from the mainstream beliefs. I would also like to thank my supervisory committee: Dr. Mikhail Raikh - who always believed in me and encouraged me to continue with my Ph.D., Dr. Christoph Boehme for his advice and lab resources, Dr. Carlton DeTar, and Dr. Michael Morse. The help and instructions provided by Dr. Randy Polson, Dr. Matt DeLong, Mr. Leonard Vojcik, and Wayne Wingert were invaluable. I would like to acknowledge Dr.Vladimir Burtman, Dr. Tromer Drori and Dr. Alexander Ndobe who introduced me to the field and the experimental techniques. I would like to thank Drs. Darin Laird and Sergey Li from Plextronics for their collaboration, support and invaluable new materials synthesized at Plextronics, Inc. I would also like to thank my research group members: Dr. Tho Nguyen, Bhoj Gautam, Bill Pandit, Tek Basel, and Ella Olejnik for the collaboration, useful ideas, and fruitful discussions. My dear friends Maria Navas Moreno and Su Liu deserve special thanks for the emotional support during these long five years. It is impossible to sufficiently express my gratitude to my husband Richard Paul Markosian and my family for their continuous support, patience, and love during my studies. x Finally, I would like to thank the funding agencies DOE (grant No DE-FG02- 04ER46109) and NSF (grant No DMR 08-03325). CHAPTER 1 INTRODUCTION For a long time all polymers have been regarded as insulators. Moreover, any electrical conductance in polymers, which was mostly caused by loosely bound ions, was considered an undesirable effect.[1] However, in the 1970s significant work was done on conduction in -conjugated polymers. Consequently, in 1977, Alan J. Heeger, Alan MacDiarmid and Hideki Shirakawa reported high conductivity in oxidized iodine-doped polyacetylene. Later, they received the 2000 Nobel Prize in Chemistry "For the discovery and development of conductive polymers." Subsequently conducting polymers and their applications have become the focus of worldwide research. The flexibility, lightweight, and relatively inexpensive fabrication made polymers very attractive for a number of applications. In the past 30 years significant progress was made in the field of organic photovoltaics [2-6], organic light emitting diodes [7], organic field effect transistors [8], organic spin-valves [9], biological sensors, etc. In this chapter we will briefly introduce π-conjugated polymers and their electronic processes; then we will introduce some of the applications such as organic 2 solar cells, organic light emitting diodes and organic spin-valves. In addition, we will mention some of the discoveries in organic spintronics and introduce magnetic field effect. The focus of this work is mainly to study organic materials that can be used in organic photovoltaics as donor and acceptor materials. Using steady state optical and magnetic resonance probes we examine electronic processes in these materials. We also compare several donor and acceptor materials to understand how the morphology and the chemical structure can affect their optical and magnetic resonance properties. All of this can bring us closer to the understanding of why some materials are better suited for use in organic solar cells. 1.1. П-conjugated polymers The term polymer refers to a large molecule consisting of a number of repeat units connected by covalent chemical bonds. Most commonly, the continuously linked backbone of the polymers studied here consists of carbon atoms. The carbon atoms in the polymer backbone can be connected by single (sp3 hybridization), double (sp2 hybridization), or triple bonds (sp hybridization) as shown in Figure 1. If all carbon atoms are connected by single bonds, the result is a saturated polymer that behaves as an insulator. In -conjugated polymers, the sp2pz hybridization causes one electron to be bound in π bond (pz) perpendicular to the polymer backbone (sp2), while the other three electrons form planar -bonds with the two neighboring intrachain carbon atoms and one side group. The π-electrons are delocalized over a number of intrachain C-atoms [10]. 3 Figure 1 Hybridization types of carbon atoms and bonds in conjugated polymer subunits. a) ethane with single bonds and sp3 hybridization, b) ethylene with a double bond and sp2 hybridization, and c) acetylene with a triple bond and sp hybridization Since there is one π-electron per C-atom in the chain, these types of polymers can be viewed as one-dimensional metals with half-filled electronic band. One-dimensional metals distort and dimerize due to electron-phonon coupling [11]. For t-(CH)x in the dimerized configuration, the polymer backbone is made of a series of short double bonds and long single bonds. This causes formation of an energy gap at the metal Fermi energy that was named the optical gap. For other polymers benzene rings or five-member rings are also involved in the dimerization; these ministructures enhance the energy gap due to electron-phonon interaction, by adding a structural component to the gap equation. 1.1.1. Model for stable excitations in -conjugated polymers Su, Shriefer, and Heeger proposed a model, named SSH model, for t-(CH)x based on the tight binding approximation, taking into account only electron-phonon interaction, but ignoring electron-electron interaction [12]. In this model they applied a semiclassical Huckel Hamiltonian that consists of two components: the classically treated lattice kinetic energy and quantum mechanically treated electron-phonon interaction as seen in equation (1.1): 4 , + +1, , (1.1) where t0 is the hopping integral between the nearest carbon neighbors on the undistorted chain, α is the electron-lattice coupling constant, C+ n,s and Cn,s are creation and annihilation operators of an electron on a site n with a spin s, K is spring constant due to π-electrons, and un is deviation of the n-th side from the equilibrium position in an undistorted chain with equal distances between the sites. According to this model the dimerization lowers the system energy and opens the energy gap, Eg = 2, where Δ=4αu, and u is the dimerization amplitude in equilibrium. Consequently, the occupied electronic states in the equilibrium are lowered resulting in a more stable configuration. Therefore, the system has properties of a semiconductor instead of a one-dimensional metal, since the density of states at the Fermi level is null. Models that take into account the electron-electron interaction and tri-dimensional chain coupling are based on the Hubbard Hamiltonian. (1.2) where U is the Coulombic repulsion between two electrons on the same site, ni,↑ and ni,↓ describes the density operators for electrons with spin up and spin down. However, the Hubbard model does not take into the account electron-phonon interaction. Therefore, more realistic model would be a combination between the SSH and Hubbard model. This is the Pariser - Parr - Pople (PPP) model, for example [13] 5 1.2. Major photoexcitations Two types of photo-excitations are dominant in the π-conjugated polymers: neutral (excitons) and charged (polarons) photo-excitations. After photon absorption neutral, spinless excitons are photogenerated that later may either dissociate into charged excitations or undergo intersystem crossing to other type of neutral excitation with non-zero spin (Figure 2). These photoexcitations will be discussed in the following sections. 1.2.1. Excitons Excitons are electron hole pairs bound through their mutual Coulombic interaction. They can have kinetic energy. In the dipole approximation excitons are formed via absorption of a single photon when electron is promoted to a higher energy level. The excitation causes structural and polarization relaxation of the surrounding geometry that leads to the exciton binding energy, Eb. In most π-conjugated polymers the value of the intrachain Eb is between 0.3 - 0.5 eV. Based on the spin alignment, the electron and hole in an exciton can form singlet or triplet state with total spin 0 or 1, respectively, both excitations still remaining neutral. The wavefunction that describes the exction (two particle system) is antisymmetric in the spin and the electron coordinates and can be obtained from the Slater determinant: (1.3) where ψi(r) represent the electron part and χi(σ) represents the spin part of the wavefunction. The wavefunction with total spin quantum number S=0 corresponding to the singlet state and three wavefunctions with S=1 or triplet states are given by equations 6 (1.4) through (1.7) where the arrows ↑ and ↓ represent spin up and spin down projection of χ: (1.4) (1.5) (1.6) (1.7) The exciton bands can be seen in Figure 2. The radiative recombination or photoluminescence (PL) from a singlet state is a fast process of the order of 100 picoseconds (ps). However, it is possible for an exciton to cross from singlet into triplet manifold via intersystem crossing (ISC) within ~10 nanoseconds (ns). The ISC process is permitted due to the spin flip of one of the exciton‟s electron due to spin-orbit coupling, hyperfine interaction, or existence of radical impurities in the polymer chains. The optical transition from the triplet excited state to the singlet ground state is ordinarily forbidden. However, in π-conjugated polymers a weak radiative transition from the excited triplet lowest state, or phosphorescence (PH), is sometimes observed. This transition is allowed due to spin flip of one of the electrons caused by a spin-orbit interaction. This transition is relatively long-lived on the order of several milliseconds (ms) [14]. 7 Figure 2 Photoexcitations in conjugated polymers: polaron excitation with P1 and P2 transitions on the left and exciton bands on the right. The triplet state has a lower energy than the singlet state. In CW photoinduced absorption we only can see long lived photoexcitations with the lifetime of the order of one millisecond such as absorption due to triplet exciton and polaron: P1 and P2 transitions. However, we cannot detect PA1 and PA2 transitions in the singlet manifold because of their short lifetimes. 8 1.5.1. Polaron excitations The intermolecular forces in organic materials are relatively weak resulting in a less rigid structure compared to inorganic materials such as silicon. Consequently, a moving charge carrier can cause local distortion and form a charged quasiparticle with nonzero spin dubbed polaron. A polaron can be charged positively (P+) or negatively (P-) and has spin ½. Polarons have two localized states in the gap and usually two allowed optical transitions P1 and P2 as shown in Figure 3. The third transition from u to u, or from g to g, is forbidden, but can become somewhat allowed due to symmetry breaking on the chain, for example by strong interchain interaction, or the existence of a sulfur atom with strong bonding. Polarons move from chain to chain through hopping, and they are the major charge/current carriers in organic device applications such as organic photovoltaic devices, as will be discussed later in this chapter. 1.5.1. Polaron pairs A polaron pair shown in Figure 4 is formed by two oppositely charged polarons on adjacent polymer chains [15]. A polaron pair that is generated by relaxation of a higher energy singlet exciton is called a geminate pair and it has spin 0. If a polaron pair is generated via an unrelated electron and a hole, then it is called a nongeminate polaron pair, and it can have spin 0 or 1 with a higher probability of having triplet than singlet configuration (3:1). 9 Figure 3 Energy levels and optical transition of positively and negatively charged polarons in the symmetric approximation. 10 Figure 4 Polaron pair energy levels and optical transitions. 1.5.2. Bipolaron excitations Bipolarons are formed by two equally charged polarons on the same site. They can be formed by either two positive polarons (BP+) or by two negative polarons (BP-). Bipolarons have two electronic levels in the gap, but only one allowed optical transition below the gap, as shown in Figure 5 [16]. BP excitations are formed in cases that there are polarons with high density; this is needed to overcome the Coulomb repulsion of two like charges on the same lattice distortion. 11 Figure 5 Bipolaron energy levels and allowed optical transitions. 1.3. Organic solar cells Conventional solar cells, or silicon p-n junction cells were invented in the 1950s [17]. Since then there were considerable advances in efficiency, reliability and cost reduction of these photovoltaic solar cells. However, the cost of solar power still remains too high. Therefore considerable effort has been made in developing solar technology based on less expensive organic semiconductors [18]. 12 The fundamental difference between organic and inorganic photovoltaic solar cells is that light absorption in the former results in the formation of excitons (coulombically bound electron-hole pair) rather than free electron-hole pairs as in the latter semiconductors. The lifetime of an exciton in organic photovoltaic (OPV) device is ~1ns after which the exciton recombines emitting light. Therefore the achievement of a successful dissociation of excitons into free charge carriers-- namely electrons and holes that can be collected at the electrodes-- has greatly influenced the architecture of organic solar cells. 1.3.1. Single layer OPV device A single layer architecture is the simplest. It consists of an organic semiconductor sandwiched between two electrodes with asymmetric work function (Figure 6). However, since most of the photogenerated excitons do not dissociate, this kind of OPV device has very low power conversion efficiency. 1.3.1. Bilayer OPV device The bilayer OPV device consists of one layer of an electron donor (for example polymer) and the other layer of electron acceptor (for example fullerene molecule) sandwiched between two electrodes as shown in Figure 6. 13 Figure 6 Processes in a single layered photovoltaic device in the rigid band approximation. One of the electrodes (ITO on the glass) must be transparent so that light can penetrate and photoexcite the polymer. Following light absorption the photogenerated exciton diffuses through the polymer and relaxes into the loosely bound polaron pair if it reaches the donor/acceptor interface. Due to the electric field at the interface the polaron pair may further separate into a free hole and electron, because the electron tends to transfer to the acceptor layer that has a lower LUMO than the polymer, whereas the hole stays on the polymer. 14 Figure 7 The schematic representation of a bilayer OPV device. The dissociation efficiency of the polaron pair at the interface is nearly 100% because it is a very fast process on the order of femtoseconds (fs), much faster than the competing processes such as photoluminescence (~ns) and recombination due to charge back transfer (~μs). Once separated, the electrons and holes may be transported to their electrodes to generated current and photovoltage. The major flaw of this architecture lies in the difficulty in finding the optimal thickness of organic materials. The diffusion length of an exciton is ~10 nm, and thus if the polymer layer is too thick, then the photogenerated exciton recombines before reaching the donor/acceptor boundary. This results in a loss of photocurrent. However, if the polymer layer is too thin for collecting all the photogenerated excitons, then the 15 absorption will be low because most of the photons will pass through the material without photogenerating excitons. 1.3.2. Bulk heterojunction OPV device The above mentioned thickness problem of the polymer layer is solved by changing the architecture. In bulk heterojunction the active layer consists of a mixture of the donors and acceptors as shown in Figure 8. Consequently, the donor and acceptor domains are spontaneously formed in the film cast from this mixture [19, 20]. The size of these domains is comparable to the exciton diffusion length, so that an exciton can reach the interface before it recombines. This results in a reduction of the exciton loss observed in bilayer devices. However, for successful transport of free charges to the electrodes, continuous pathways of acceptor (fullerene) and polymer (donor) to the electrodes are needed. Unfortunately, this is not always the case, since in a polymer/fullerene mixture the fullerene domains can be completely surrounded by a polymer (and vice versa), which results in loss due to charge back-transfer recombination. 1.3.3. Power conversion efficiency Power conversion efficiency of an OPV cell presents the percentage of the solar energy shining on the device that is converted into electricity. It can be calculated from the I-V characteristics of a solar cell, as shown in Figure 9 and equation (1.8). (1.8) 16 Figure 8 Bulk heterojunction architecture of an OPVC Jsc is the short circuit current density. Voc is the open circuit voltage. FF or fill factor is the ratio of the maximum power delivered by solar cell (illustrated by light gray rectangular area in Figure 9 and product of Jsc and Voc). Ps is the optical irradiance of the incident light from the sun, 100mW/cm2. From equation (1.8) we can see that by increasing the fill factor, Jsc, or Voc we may raise the power conversion efficiency of a solar cell. Depending on donor and acceptor materials used in organic photovoltaic cell, the power conversion efficiency may vary. In this work we present our studies of pure polymers, fullerene derivatives, as well as their mixture using optical and magnetic resonance experiments. 17 Figure 9 I-V characteristic of an organic solar cell with RR P3HT/PCBM as the active material. This measurement was done by Dr. Ye Zhang. 1.4. Organic light emitting diodes A light emitting diode (LED) is a semiconductor light emitting source. LED was introduced as a practical electronic component in 1962.[21] An organic light emitting diode (OLED) is an LED in which the emissive electroluminescent layer is an organic semiconductor film which emits light when electric current is applied to the electrodes. The organic semiconductor is sandwiched between two electrodes (Figure 10). In a more sophisticated OLED the organic material consists of an electron and hole transport layers, named ETL and HTL, respectively. The electrons and holes are injected from the opposite electrodes when a biasing voltage is applied to them. They migrate through the ETL and HTL via hopping in the form of charge polarons. The oppositely charged 18 Figure 10 The configuration of a typical OLED. ITO coated glass acts as a transparent anode and calcium as a low work function cathode. Aluminum is evaporated on the top of calcium as a protection layer. PEDOT presents hole transport layer and MEH-PPV active organic material. polarons sometimes meet in the active material, and if they are within capture distance (equation(1.9)), then they form bound polaron pairs, which are precursors to excitons. Subsequently, the exciton may recombine radiatively and emit light. The emitted light or electroluminescence (EL) is collected through the transparent electrode. (1.9) The recombination rate in the OLED depends on charge density and mobility. The recombination zone within the active layer can be adjusted by changing the work-functions of electrodes, adjusting the HOMO-LUMO levels of transport media, and changing the thickness of electron and hole transport layers. 19 The electron-hole pairs can be either in singlet or triplet configuration. However, in polymers only the singlet configuration results in emission of light. In small molecules with large SOC it may be that PH emission prevails. Because of the triplet/singlet spin statistics (3:1) these types of OLEDs are more effective. Consequently OLEDs have been categorized according to emissive materials as two different types: small molecule and polymer OLEDs. One of the advantages of polymer OLEDs that are studied in our group is that the active layer does not have to be evaporated like small molecules, but in contrast it can be spin-coated from the polymer solution. This type of processing is also suitable for large area OLED such as in TV screens. White organic light emitting diode or WOLED refers to an organic light emitting diode emitting white light. The first WOLED was reported in 1994 by Kido et al.[22] One of the challenges to be overcome is the emission spectrum. The perfect WOLED should emit continuous spectrum covering the entire visible range (400 nm to 800 nm) and should have similar spectral distribution to the sun light. The most common approach to solution of this problem is to combine electroluminescence from several lumophores (2 or 3). The lumophores can be blended into a single layer, separated into different layers in the same device, or contained in the several independent devices where each device emits light in a different color. [23] The overall efficiency of an OLED (or WOLED) depends on a number of factors such as charge recombination efficiency; the ratio of generated singlet and triplet excited states; the internal electroluminescence quantum efficiency; and the fraction of photons extracted from the structure. The charge recombination efficiency in most cases 20 approaches 1. In contrast the other factors especially singlet to triplet ratio present significant loss in the overall efficiency. According to spin statistics, 25% of all generated excitons are in singlet configuration and 75% are in the triplet configuration. Since only singlet excitons recombine radiatively in polymers, this puts an upper limit of 25% on the overall efficiency of an OLED based on these organic materials. However, this limitation can be overcome by using a heavy metal ion in a phosphorescent material[24]. A heavy metal ion relaxes the spin conservation law and promotes intersystem crossing between the triplet and singlet states. In this way it is possible to harvest emission from triplet state and increase the efficiency of an OLED. However, the successful fabrication of WOLED based on phosphorescent materials is complicated because of fast degradation and instability of blue phosphorescent emitters. Therefore, the focus shifted to fabricating WOLEDs with luminescent blue emitters and phosphorescent emitters at longer wavelengths. Consequently, it is possible to recycle triplet states from blue luminescent emitters by transferring them to phosphorescent emitters at longer wavelengths where they can decay radiatively.[23] 1.4.1. Magnetic field effect As mentioned earlier in an OLED device the generated polaron pairs can be either in singlet or triplet state. In the absence of an external magnetic field, triplet states are degenerate, and the singlet PP mixes with an entire PP triplet manifold due to the hyperfine interaction. However, when a strong magnetic field is applied to the device the degeneracy of the triplet states is lifted due to the Zeeman interaction, and then singlet PP 21 mixes only with the triplet PP spin sublevel that has magnetic quantum number 0. Consequently a magnetic field can influence the formation of singlet and triplet excitons depending on their generation rates. Increased formation of singlet excitons at the expense of triplet excitons would certainly benefit the OLED overall efficiency. 1.5. Organic spintronics For the past decade the potential of organic spintronics has been recognized. Organic semiconductors can preserve the spin information over extremely long times, and therefore they can be used to carry information by means of the electron spin. In order to carry information via spin, three basic processes must take place. First, the spin of electron population must be polarized. Second, the spin polarized electrons must diffuse through the material. And finally the analysis of the spin polarization is conducted. The degree of spin polarization indicates the spin-relaxation time or the time required for the spin direction to change because of interaction with environment. The spin relaxation is affected by spin-orbit coupling and hyperfine interaction. Since both spin-orbit coupling and hyperfine interaction in organic semiconductors are relatively weak, the degree of polarization can be found through measuring of magneto-resistance of a spin-valve, where organic material is placed between two ferromagnetic electrodes.[9] 1.5.1. Spin-orbit coupling The spin-orbit coupling is the interaction between the electron spin and its orbital motion around atomic nucleus. Spin-orbit coupling occurs when a particle with nonzero spin moves in an electric field. Even though there is no magnetic field in the restframe of 22 the nucleus, the electron, because of its velocity, feels a magnetic field, as shown in equation (1.10). . (1.10) From here we can see that the spin of an electron can be indirectly influenced by an electric field. For hydrogen-like atoms spin-orbit interaction is proportional to Z4. Since organic semiconductors have low atomic numbers, the spin-orbit coupling in them is generally small. 1.5.2. Hyperfine interaction The hyperfine interaction is an interaction between electron spin and nuclear spins of the host material. Since the electron can interact with more than one nucleus, the hyperfine interaction can be described by the Hamiltonian expressed in equation (1.11). (1.11) where Ai is a coupling intensity between the spins, Ii is the nuclear spin operator and S is the electron spin operator. Nuclear spins can also affect the spin relaxation and cause dephasing. For example, if an electron spin interacts with N nuclear spins, the statistical fluctuation scales with 1/√N. From here we can conclude that as an electron interacts with more nuclei, or as it is more delocalized, the weaker the influence of nuclei on spin relaxation. 23 The nuclear spins in organics mostly originate from isotopes such as 1H (I=1/2), 13C (I=1/2), or 14N (I=2). Consequently, the hyperfine interaction as well as spin relaxation time can be altered when an organic material is enriched with these isotopes, as it is the case with deutereted DOO-PPV discussed in Chapter 6. . CHAPTER 2 EXPERIMENTAL TECHNIQUES This chapter introduces the experimental techniques applied in characterizing films of polymers, fullerene derivatives, or their blends that can be used in organic photovoltaics. Thin films are prepared using glass, sapphire, or cesium-iodide substrates by either drop casting or spin coating in inert nitrogen atmosphere inside a glove box with oxygen level less than 1ppm. 2.1. Absorption and emission The absorption of light by conjugated polymers promotes an electron from a ground state S0 to a higher singlet electronic state S1 that is coupled to the ground state. Transitions from the ground state to higher electronic states Sn can occur where n represents higher laying state with appropriate parity, angular momentum and oscillator strength with respect to ground state. Absorption in visible, ultraviolet and near infrared range is measured by using CARY 17 spectrophotometer. The film is deposited on glass substrate. In order to eliminate substrate effects and the system response, absorption of blank substrate is measured first and then automatically subtracted from the absorption spectrum of the sample. Scattering and reflection from the films are assumed to be small and therefore are not accounted for. Absorption is measured in units of optical density (OD), which is 25 related to transmission by T=T010-OD, where T0 is transmission of the system without the film and OD=αd, where α is absorption coefficient and d is the thickness of the film. One of the excited states decay channels is radiative decay or photoluminescence (PL) where exciton relaxes to ground state S0 by emitting a photon with appropriate energy ћω, corresponding to the optical transition. In order to measure emission (photoluminescence) we use CW laser with a minimum energy of S1-S0 optical transition, this way the excitons can be photogenerated in the film. The emission light is collected by a spherical mirror and dispersed by a grating monochromator enabling measuring of the spectrum. Dispersed light is detected by an appropriate solid state detector depending on wavelength range, such as Silicon photodiode (ћω >1.1eV) or germanium detector (ћω>0.8eV). Photoluminescence spectra reported in this diesertation are done under vacuum and at low temperatures. 2.2. Continuous wave (CW) spectroscopy CW spectroscopy is a powerful tool to study long lived species such as triplet and polaron excitations. 2.2.1. Photoinduced absorption Photoinduced absorption (PA) is the experimental technique used to study long lived species (with life time of the order of 1ms at low temperatures) such as triplet and polaron excitations.[25-27] Ar+ ion laser is used in most cases as a pump to promote electrons from the ground to an excited state. Tungsten or Xenon lamps are used as probe light covering the wavelength range from 250nm to 4.2μm. Pump and probe beams are overlapped on the sample placed in vacuum inside of a cryostat. The light coming out of 26 the sample is collected in one of the spherical mirrors and focused onto 2mm wide slit of monochromator. The appropriate long pass filters are placed at the slit of monochromator to block the pump beam. The light is then collected at the exit of the monochromator with a suitable photodetector: Si 10D photodiode for wavelengths from 550nm to 1.05μm, Ge detector from 800 to 1.6μm, and InSb from 1μm to 4.2μm. The signal is then converted from current to voltage and amplified via preamplifier and sent to lock-in SR 830 amplifier. In order for the signal to be detected by the lock -in amplifier the pump light has to be modulated with a frequency corresponding to the lifetime of photoexcitation, usually around 300Hz. This experimental set-up is presented in Figure 11. The absorption of the photoexcited species is essentially the difference in transmission (ΔT) when the sample is illuminated with both the pump and the probe (TL) and when sample is not illuminated with the pump but only with the probe (TD). The relation between the induced absorption coefficient (Δα) and difference in transmission (ΔT) is expressed the equations (2.1) through (2.4): Δ (2.1) 1+ ΔT TD = e -Δαd (2.2) α Δ (2.3) Δ Δα Δ (2.4) 27 Laser - pump Mirror Mirror Mirror Chopper Light source - probe Sample in Cryostat Monochromator Chopper controller Band pass filter Detector Pre amplifier Lock in amplifier Signal Reference Figure 11 The photoinduced absorption experimental apparatus Here we can distinguish between two signals: photoinduced absorption (PA) when Δα < 0 and photoinduced bleaching (PB) or depletion of lower energy state involved in photonic transition, when Δα > 0. 2.2.2. Doping induced absorption In doping induced absorption (DIA) the change in transmission, ΔT of the film‟s optical transmission, T induced by doping is measured vs. photon energy using phase sensitive technique. The end result is the absorption-modulation spectrum, ΔT/T. In this experiment the films were doped by exposure to iodine (I2) vapor for several seconds (~20s). First the transmission of undoped sample, T is measured. Then 28 the transmission of doped sample, TD is measured. The ΔT/T spectrum is obtained by subtracting the transmission after doping from transmission before doping, and then normalizing it by transmission before doping. In a one-dimensional model single charge carrier added to a polymer chain forms a spin ½ polaron with two localized states in the energy gap. As mentioned in the previous chapter polarons have two allowed optical transitions P1 and P2. In this experiment large iodine ions may isolate polymer chains and cause quasi one-dimensional electronic properties. Furthermore, the iodine counter-ion may localize the induced polaron excitation. As a result a DIA spectrum consists of two polaron bands at low and high photon energy, P1 and P2. 2.3. Modulation spectroscopy Modulation spectroscopy is used to understand the kinetics of the photogenerated species. The pump is modulated at a frequency f with a mechanical chopper or an acousto-optic modulator. PA is measured by lock-in amplifier tuned to the modulation frequency, f. The lock-in amplifier analyzes the first harmonics of the photoinduced signal defined as G(t). (2.5) g represents the generation rate proportional to the pump intensity and ω=2πf. Two phase components of the signal are detected by lock-in amplifier: in-phase and out -of- phase or quadrature. The in-phase means that signal- response is in phase with modulation 29 frequency ω. The out-of-phase component is π/2 relative to ω representing long-time response with respect to 2π/ω. The in-phase component can be written as nI(t). (2.6) The out-of-phase or quadrature is defined as nQ(t). (2.7) (2.8) τ is the mean life time of a photexcitation. 2.3.1. Mono- and bimolecular recombination Photoinduced absorption, PA (Δα) is proportional to excitation density N(t). Therefore to determine lifetimes and type of recombination, PA as a function of a modulation frequency, ω and laser intensity, IL has been used. There are two relevant recombination mechanisms, monomolecular (MR) and bimolecular (BR) processes. Monomolecular kinetics points to a possible geminate generation, where the hole and electron polarons (P±) originate from a common excitation parent. Bimolecular kinetics, on the other hand, can indicate a distant pair recombination. [28] These processes can be described by a single rate equation for the photoexcitation density: (2.9) 30 where G(t) is proportional to generation rate g: (2.10) g is proportional to pump intensity IL: (2.11) αL and η represent the absorption coefficient at laser photon energy and a quantum efficiency in respective order. Equation (2.11) includes S(N0) function that includes defect limited cases in which due to trap filling N tends to saturate at high I. N0 represents the trap density and U recombination mechanism.[28] However, for simplicity we assume that S=1 and equation (2.11) becomes (2.12) where for MR and for BR and In CW measurements we assume steady state conditions, when (2.13) In the case of monomolecular recombination the equation (2.15) becomes 31 (2.14) with a solution: (2.15) According to the equation (2.17) if monomolecular recombination occurs, the PA signal depends linearly on laser excitation intensity. However, in the case of bimolecular recombination the recombination rate is proportional to population squared as mentioned above , where b is proportionality constant. The equation 2.16 becomes =0, (2.16) (2.17) In the bimolecular recombination the PA signal depends sublinearly on laser intensity. 2.3.2. Modulation frequency response The dynamics of long-lived photoexcitations can be successfully studied by looking at their PA dependence of modulation frequency and the intensity of the pump. In the case of finite modulation frequency, the in-phase and quadrature components can be measured using phase sensitive lock-in technique and are given by equations (2.18) and (2.19)[28]. 32 (2.18) (2.19) If we replace τ with τBR=√(gb) we get expressions for bimolecular recombination. NQ equals zero at zero frequency and has a maximum at (2.20) where NI(ωmax)=NQ(ωmax). At higher modulation frequencies ω> ωmax , both in-phase and quadrature component exhibit asymptotic behavior with a power law decrease ω-β, where β=2 for in-phase and β=1 for quadrature component. In this measurement is very important to take into account the system response and to phase the measurement results relative to the laser phase. In order to do that, before each measurement the laser should be scattered at monochromator slit without any optical filters while the probe is blocked. The lock-in amplifier should be auto-phased so that the signal is completely in the in-phase component. Since the measurement consists of collecting in-phase and quadrature components while with help of acusto-optic modulator the modulation frequency is changed, signal tends to get out of phase in respect to the pump. This offset can be accounted for in the following way: Complete frequency scans, both in-phase and out-of-phase are taken of PA, PL, and of the modulated laser (Lin(f) and Lout(f)). Then, 33 (2.21) Δ (2.22) Δ (2.23) Δ Δ (2.24) Δ Δ Δ Δ Δ (2.25) R(f) is modulus of measured PA and α(f) is relative phase between the measurement, υΔA and the modulated laser response, υL. Based on equations (2.18) through (2.22) we can calculate actual PA frequency response for both in-phase and out-of-phase components. (2.26) (2.27) 2.3.3. Dispersive kinetics The equations presented above can be used to describe kinetics in ordered materials. However, in this study we concentrate on polymers which are disordered materials so most processes are temporarily dispersed. Therefore the recombination processes can be described more accurately by the simple Cole-Cole model for amorphous materials [29]: 34 (2.28) where α is dispersive parameter α<<1, and τeff is mean lifetime. The in-phase and quadrature are given by NI = Re(N) and NQ = Im(N). In ordered material photoexcitations have a single lifetime and α=1. In polymers (disordered medium) dispersive parameter, α is less than 1 and τ becomes τeff. PA response at high modulation frequencies distinguishes dispersive kinetics from the single lifetime processes. Figure 12 shows PA dependence from modulation frequency for a polymer/fullerene blend. Using equations (2.23) through (2.29) to process collected data and fitting with equation (2.30) we can find the lifetime of the photoexcitaion and the recombination mechanism.For dispersive kinetics, described by Equation (2.30) the in-phase and the quadrature components decrease sublinearly with ω: , where β ≈ α. 2.1.1. FTIR photomodulation spectroscopy Fourier transform infrared (FTIR) spectroscopy is used to measure photoinduced absorption in mid and far infrared photon energy range. This technique is based on Michelson interferometer, depicted in Figure 13. Beam I0 coming from an appropriate light source is split by a beam splitter into two, I1 and I2. Beam I1 is reflected from a stationary mirror and I2 from a movable mirror causing the change in optical travel length denoted by x. The two reflected beams pass through the sample generating an interferogram spectrum that can be described by the equation (2.29). 35 Figure 12 Frequency dependant PA of LYON:PC70BM blend. 36 Figure 13 Schematics of Michelson interferometer (2.29) where Iout(x) is interferogram spectrum as a function of displacement x. The inverse Fourier transform of the second part of Equation(2.31) describes the interference spectrum as a function of wavenumber , given by (2.30) 37 FTIR spectroscopy uses the fast Fourier transform (FFT) to convolute spectrum with a limitation on resolution, Δνmax. Here FTIR spectroscopy is used to construct PA spectrum in range of mid- to far- IR, so that: Δ (2.31) where Ton and Toff are transmitted spectrum with and without illumination by the pump laser. Ar+ ion laser was used as the pump. The laser beam was directed through a shutter that is controlled by the computer. The globar lamp with emission in IR spectrum was used as a probe. The IR beam was split by KBr beam splitter that is fit to a range 400 cm- 1 to 5000 cm-1.The Light beam that passed through the sample is collected by a MCT (mercury-cadmium-telluride) detector that must be cooled with liquid nitrogen. The FTIR experimental setup is shown in Figure 14 2.1. Electroabsorption In electroabsorption or EA measurement an electric field is applied on the sample and then the change in transmission is measured with the probe similarly to previously described PA experiment. Transmitted light intensity ΔI/I decreases exponentially with absorption constant α and film thickness d. Therefore, change in absorption constant, Δα can be derived from the change in transmittance: Δ Δ (2.32) 38 . Figure 14 The photomodulation experimental set-up using FTIR spectrometer Since in the electroabsorption experiment electric field is applied, a sample/film was deposited on a special substrate prepared in the following way: A 0.5nm titanium film followed by 150nm gold film was RF sputtered on a 1 inch in diameter and 2mm thick sapphire substrate. Then an interdigitated gold electrode array was photo lithographically patterned and etched with a 40μm gap between the electrode fingers as shown in Figure 15. 39 Figure 15 Electroabsorption substrate architecture. Figure 16 illustrates the EA experimental set-up. The applied electric field was produced using signal generator (Tekttronix -FG503, 3MHz) and transformed using a step-up transformer. The deposited film on the EA substrate was supplied with 200-300V AC that generated an AC field of the order of 105V/cm at frequency of 500Hz. Depending on desired wavelength range either xenon or tungsten lamp was used as a probe light. Probe light was dispersed through a grating monochromator and guided with mirrors through the sample into a solid state detector. The sample holder was mounted on 40 Figure 16 Schematic representation of EA experimental set-up a cold-finger cryostat and its electrodes connected to the leads from the transformer. Using a phase sensitive modulation technique, a reference signal is taken from the signal generator and the signal from the detector was connected to the input of the lock-in. Because conjugated polymers have mirror symmetry, they respond to a sinusoidal field of frequency with signals at 2ω, where fundamental frequency is significantly suppressed. Nonzero signal at fundamental frequency indicates existence of asymmetry, possibly due to internal fields. Therefore, an EA signal was detected at a second harmonic frequency, indicating transitions due to the change of electronic states. 41 2.2. Photoluminescence quantum efficiency Photoluminescence quantum efficiency (PLQE) can be defined as number of emitted photons per absorbed photon. PLQE provides useful information for better understanding of radiative and nonradiative processes in organic materials. The measurement presented in Figure 17 is performed using Ar+ ion laser as a source of excitation and integrated sphere (IS) that distributes light homogenously. The sample is placed inside the sphere and the light emission is collected at the exit of IS with a solid state detector. The signal is amplified and analyzed using lock-in technique. Before starting the actual PLQ E measurement optical density (OD) of the thin spincoated film is measured. The OD should be below 1. The PLQE can be calculated using the equation (2.33). Figure 17 PLQE experimental set up done at room temperature in air 42 (2.33) IL, IPL and IPL,corr are measured and represent: the reflected laser intensity without the sample inside IS, the uncorrected PL intensity with sample inside the laser path, and the corrected PL intensity with the sample inside IS but out of the laser path. Correction term, IPL,corr is measured to eliminate the contribution of reflected photons (from IS walls) to collected light emission. R and T are reflection and transmission coefficients at laser wavelength. DL and DPL are detector sensitivities at laser and the PL wavelengths. SL and SPL are IS sensitivities at laser and PL wavelengths. EL and EPL are photon energies of laser and PL emission. TF(PL) is transmission of optical long pass filter used in the experiment at PL wavelength. 2.3. Electron spin resonance spectroscopy 2.3.1. Introduction to theory The physics of electron spin resonance is the basis for two major experimental techniques used in this study, ESR and ODMR. 2.3.1.1. Spin and orbital moments Electron spin resonance (ESR) studies the interaction between electronic magnetic moments and magnetic fields. The magnetic moment of an electron spin, μs and magnetic moment associated with its orbital momentum μL can be expressed as: 43 (2.34) (2.35) where S is the spin angular momentum operator, L is the angular momentum operator, and, β is Bohr magneton, defined by (2.36) If an electron has a both spin and orbital motion, then the total angular momentum J is obtained by the vector addition as in equation (2.37). (2.37) where J has possible magnitudes |L-S|, |L-S+1|,…,|L+S|. As a result of the vector addition of the spin and orbital components the overall magnetic moment can be expressed as: (2.38) where g is the Landѐ factor . (2.39) In the solids the electrical orbital motion interacts strongly with the crystalline electric fields and becomes decoupled from the spin or quenched. The stronger the 44 quenching the closer the g factor to the free electron value, 2.0023. [30] The value of g factor depends on the environment of the unpaired electron. It depends on the orientation of the molecule containing the unpaired electron in respect to applied electric field, as well as the phase of the sample (solid, liquid, or gas). In the gas and liquid phase the molecules have free motion and value of g is averaged over all orientations. In the case of paramagnetic ion inside of a perfect cubic crystal the g value does not vary with orientation. [31] However, in most cases such as a radical ion in a crystal with low symmetry, the g factor is very often anisotropic and varies with the direction (x‟,y‟z‟) in a single crystal. Generally, the g factor may be a symmetric tensor g' with components g‟ij (2.40) where g‟ij=g‟ji. However, it is possible to find principal axes (x,y,z) where the g tensor is diagonal. (2.41) Very often the g tensor has axial symmetry where (2.42) 45 (2.43) where the z axis is taken as the symmetry axis. For an arbitrary orientation of a crystal in a magnetic field, the resonance is characterized by the g factor (2.44) where θ represents the angle between an axis and direction of magnetic field. Since the three direction cosines obey the relation (2.45) Using equation ( 2.45) and assuming that θ is the angle between the symmetry axis and magnetic field equation ( 2.44) can be simplified to (2.46) 2.3.1.2. Spin Hamiltonian The interaction energy of a paramagnetic atom in a constant magnetic field H0 is represented by the spin Hamiltonian : (2.47) 46 represents electronic energy of the paramagnetic ion in free state, crystal field energy or interaction energy of the free ion‟s electronic structure with the crystalline electric field , and spin orbit coupling, spin-spin interaction, Zeeman energy, hyperfine structure, quadrupole energy, and nuclear spin energy. However, in our studies of ESR, LESR and ODMR the most important components are Zeeman, hyperfine interaction and to some degree spin-spin and spin-orbit interaction. Spin-orbit interaction is defined by the equation (2.48). (2.48) where λ is the spin-orbit coupling constant. The spin-spin interaction is defined by the equation (2.49). (2.49) where D is the zero field splitting parameter and Sz is the z component of the spin angular moment. is Zeeman energy (2.50) 47 Hyperfine structure is given by (2.51) where Ax, Ay, and Az are components of hyperfine coupling constant A, and Ix, Iy, and Iz are components of nuclear spin I. The ESR measurement in essence studies Zeeman energy, and the way in which the other Hamiltonian terms perturb or are perturbed by Zeeman energy. Spin orbit interaction λL∙S further splits the optical energy levels and influences g factor. The spin-spin interaction can in some cases be the same order of magnitude as the Zeeman energy causing the energy levels to be strongly dependant on orientation of crystal in magnetic field. The hyperfine structure is caused by the interaction of the nucleus with an unpaired electron. In some systems such as aromatic molecules, where electron is delocalized hyperfine structure can be expressed as: (2.52) where the projection mi of the nuclear spin on the direction of the magnetic field can have the following values: Ii, Ii-1, …, -Ii+1, -Ii. Hyperfine coupling constant Ai varies with the type of nucleus and represents the strength of the interaction between spins of nucleus and electron. Hyperfine interaction causes further splitting of energy levels (Figure 18). In the case of several nuclei with spin ½ with the same hyperfine constant A, they 48 Figure 18 Due to the Zeeman Effect in the presence of an external magnetic field the electronic energy level splits into two ±gβH energy sublevels. Each of the two electronic energy levels is split into six equally spaced energy levels by hyperfine interaction for I=5/2. Quadrupole interaction shifts these levels according to selection rule ΔmI=0 causing asymmetry. produce an ESR spectrum whose intensity follows the binomial coefficient distribution. If the width of spectral lines is smaller than their separation they are well resolved. However if their width exceeds the separation, spectral lines merge into one line. If there are n nuclei with spin I then the energy level will be split into (2I+1)n lines. If all of the coupling constants are different then no degeneracy occurs. If some nuclei have the same coupling constant then resulting degeneracy decreases the number of lines but increases their amplitude. Usually all of the hyperfine components have the same line width and shape except in the case of overlapping resonances. 49 2.3.1.3. Information obtained from ESR ESR gives us information about the various parts of Hamiltonian. The Zeeman and hyperfine structure Hamiltonian can be evaluated directly from ESR spectra. Crystal field and spin orbit energies are evaluated from optical spectra and then correlated with ESR data. In this study, ESR data are used to evaluate the g factor of various organic materials used in photovoltaic applications. ESR can be used to determine concentration of paramagnetic species. Therefore, in the studies of the blends, ESR proved to be very useful in comparing concentrations of paramagnetic species in donor and acceptor materials. In order to obtain the most of information ESR is done at different temperatures and different microwave powers. 2.3.2. ESR kinetic studies 2.3.2.1. Bloch equations In ESR studies magnetic field is applied to an electron causing spin magnetic moment to align parallel or antiparallel to direction of the magnetic field. The spins that are aligned parallel to the applied magnetic field have lower energy than the antiparallel spins. For an ensemble of spins the nonzero net magnetic moment (macroscopic magnetization) M can be expressed by the following equation (2.53) 50 In equilibrium M is in direction of magnetic field B. However if M is not in direction of B then there will be a net torque that causes M to precess about direction of B according to the equation (2.54). (2.54) In addition to precession there two relaxation processes, spin-lattice and transverse relaxation.[32, 33] Spin -lattice relaxation involves transfer of energy from the spin system to the surroundings or lattice. The time that is required for this process is called spin-lattice relaxation time, T1. (2.55) where Mz is z component of magnetization under nonequilibrium conditions, and M0 is equilibrium magnetization along B. According to the equation above Mz approaches M0 (equilibrium) in characteristic time T1. If we assume that M is not exactly in the z direction but tilted toward the x direction of the magnetic field, each individual spin will precess under a different rate. So an ensemble of spins will start a precession in phase but with time it will lose its coherence. The characteristic time for this process is called the transverse relaxation time, T2. The transverse components of magnetization, Mx, and My, decay to zero (equilibrium) according to the equation (2.56). 51 (2.56) Based on the equation (2.56) the transverse relaxation does not involve energy transfer from spin to lattice, but it does increase entropy of the system. However, the spin-lattice relaxation process where Mz goes to zero causes both transverse components to become zero. Therefore, transverse relaxation time includes both spin-relaxation time as well as spin dephasing relaxation time T2*. (2.57) Relaxation processes can be classified by their effects on electron spin. Nonsecular process is a process that involves spin flip and energy transfer between the spin system and the lattice, and therefore involves T1. Secular process does not involve spin flip but results in the loss of spin coherence. 2.3.2.2. CW ESR In an ESR experiment beside a static magnetic field B0 in z direction another oscillating magnetic field B1 in xy plane is applied. If B1 is a circularly polarized field rotating around the z axis then the total magnetic field can be expressed with the equation (2.58). (2.58) 52 Combining equations (2.54) - (2.56) we obtain equation (2.59). (2.59) Taking the vector product B × M and manipulating equations (2.58) and (2.59) we can obtain the expressions for dMx/dt, dMy/dt, and dMz/dt. However, it is very convenient to write: (2.60) (2.61) or (2.62) (2.63) In the equations (2.60) through (2.63), u represents the component of Mx that is in phase with B1 and v is 90° out of phase with B1. Differentiating these equations we obtain (2.64) (2.65) (2.66) 53 These equations express magnetic resonance in rotating coordinate frame. However, in this study we focus on the CW magnetic resonance experiment where radiation field B1 is continuous and B0 changes slowly compared to relaxation rates (slow passage condition). Therefore, we can assume steady-state conditions and set derivatives to 0. (2.67) (2.68) (2.69) where ω0=γB0 is the Larmor frequency that corresponds to the frequency of the energy level transition. According to equations (2.67 - 2.69) as B1 approaches to 0 so do u and v as well and Mz approaches to M0, the equilibrium value. In the opposite case when B1 increases Mz decreases moving away from the equilibrium. As mentioned earlier, u represents the transverse magnetization that is in-phase with B1 and therefore corresponds to dispersion. The out-of-phase response, v corresponds to the absorption. In magnetic resonance experiments, it is usually out-of-phase or v-mode detected. When microwave power proportional to B1 2 is small so that , then the equation (2.68) can be simplified. (2.70) 54 Absorption represented by the equation (2.70) corresponds to a Lorentzian line shape as seen in Figure 19. In the continuous wave electron spin resonance (CW ESR) experiments where the magnetic field B0 is changed while the microwave (MW) frequency is kept constant, half width at half maximum, ΔB0 corresponds to equation (2.71). (2.71) Figure 19 Absorption corresponding to Lorentzian line shape 55 In these studies absorption is detected via small field modulation (1G) so the signal is proportional to first derivative of absorption corresponding to first derivative of Lorentzian line shape shown in Figure 20 and represented by equation (2.72). (2.72) Here the line width can be described as a distance between two extremes of the first derivative of absorption. It can be calculated by taking a second derivative and finding zeros. Δ (2.73) Since both amplitude and line width of the absorption derivative are dependent on T2, the product of the amplitude and square of the line width (independent of T2 and proportional to M0) is taken as a measure of strength of ESR signal. In the regime of the small B1, line width is independent and amplitude is linearly proportional to B1. Neither of them is dependent on T1. However, at high microwave power the term cannot be neglected. The amplitude of ESR is not linear in B1 anymore, and both amplitude and line width depend on T1. At high microwave power amplitude starts to decrease with increasing B1. This phenomenon is known as resonance saturation. It can be understood as population equalization of energy levels that causes net absorption to decrease. 56 Figure 20 Detected ESR signal corresponding to first derivative of absorption 2.3.2.3. Line shape In a perfect scenario, the ESR signal should look like a first derivative of a Lorentzian line. However, when the magnetic field is not homogeneous and varies over the sample, not all of the molecules are in resonance at the same value of the magnetic field. This as well as unresolved hyperfine coupling might cause inhomogeneous broadening and the resonance looks more like a Gaussian line shape than a Lorentzian. In this study we work with the films of blend of polymers with fullerene derivatives and it is 57 common to see a Gaussian line shape of an ESR resonance that can be described by the equation (2.76). ω ω ω (2.74) where half width at half maximum (HWHM) can be expressed as (2.75) 2.3.3. Experimental setup for ESR ESR or electron paramagnetic resonance (EPR) technique is based on measuring the absorption of electromagnetic radiation (microwave) by an electron spin system. The sample is placed in magnetic field and transitions between Zeeman levels are monitored. Principles of ESR have been described previously in the introduction. However, the ESR measurement can be simplified by considering first only spin angular momentum S=1/2. When an electron is placed in magnetic field, the projection of its magnetic moment on the axis defined by the magnetic field, takes on two values -1/2 and +1/2. The separation ΔE between these two energy levels is linearly proportional to strength of magnetic field B. When an incoming microwave photon has energy equal to ΔE transition between the spin levels occurs as shown in Figure 21. These transitions are magnetic-dipole allowed, and are affected by the microwave magnetic field vector B1, that is perpendicular to the external magnetic field generated by the electromagnet. The sample is placed in a resonator that maximizes the B1 field at the sample. The matching of the photon energy 58 Figure 21 Zeeman energy splitting for an electron in magnetic field of the microwave, ћω, to separation of electronic levels is called resonance.[34] The simplest ESR spectrometer has three basic parts: a source of electromagnetic radiation, a sample, and a detector. In the most basic terms, the spectrum is acquired by changing the frequency of the electromagnetic radiation and measuring with a detector the amount of radiation that passes through or gets reflected from the sample in order to observe the spectroscopic absorptions. In this study we used Bruker Elyxis 580 spectrometer. The general schematic of a Bruker EPR spectrometer is shown in Figure 22. The microwave source and detector are placed inside of a microwave bridge. The sample is placed inside of a metal microwave cavity with the purpose to amplify weak signals coming from the sample. The cavity with the sample is placed inside of a cryostat because most of the measurements take place at cryogenic temperatures. The electromagnet is placed around the cryostat. The console contains signal processing and 59 Figure 22 General schematics of a ESR Spectrometer control electronics as well as a computer that is used for analyzing data and coordinating all of the units for acquiring a spectrum. 2.3.3.1. The microwave bridge As mentioned previously the most import parts in the microwave bridge are the microwave source and the detector. The microwave power that the sample sees is varied via an attenuator that blocks the flow of microwave radiation. Most ESR spectrometers are reflection spectrometers, meaning they measure changes in reflection from the microwave cavity (together with the sample) due to spectroscopic transitions.The Shotky barrier diode is used as a microwave detector. It converts microwave into current. At low power levels less than 1μW, the diode works in a square region where the diode current is proportional to microwave power. At powers larger than 1mW, the diode works in a linear region where the diode current is proportional to the square root of microwave power. To achieve reliable results the diode should work in a linear region, and the diode 60 current should be around 200μA. To ensure that detector operates in that region the reference arm supplies the detector with some extra microwave power or bias. Some of the source power is tapped off into the reference arm where the second attenuator controls the power level and consequently the diode current for optimal performance. The phase shifter ensures that reference arm microwave is in phase with reflected microwave when they combine at the detector. 2.3.3.2. The ESR cavity The microwave cavity is a metal box of rectangular or cylindrical shape that resonates with microwaves. At resonance, the microwave cavity stores microwave energy¸ no microwave will be reflected back but will remain inside of the cavity. Cavities are characterized by their quality factor or Q, which describes how well a cavity stores microwave energy. As the Q factor increases so does the sensitivity of spectrometer. (2.76) Energy can get lost due to the walls of the cavity because microwaves generate current in the walls of cavity which then generate heat. Q can be measured and calculated according to the equation (2.77): Δ (2.77) 61 where νres is the resonant frequency of the cavity and Δν is the width at half height of the resonance. Due to the resonance a standing electromagnetic wave will be formed inside of the cavity, meaning that the magnetic and electric field will be exactly out of phase. The best ESR signal and highest sensitivity of spectrometer is achieved when magnetic field is maximum and electric field as at its minimum. Microwave is coupled into the cavity via a hole called iris. With changing the size of iris it is possible to control the amount of microwave that will enter the cavity and be reflected back from the cavity. The ESR signal is basically a result of reflected microwave caused by absorption of microwave by sample and consequently lowering Q of the cavity. 2.3.3.3. Signal channel The signal channel unit located inside the console uses phase sensitive signal detection or lock-in amplifier to enhance sensitivity of spectrometer. The magnetic field strength that the sample sees is modulated sinusoidally at a modulation frequency usually 100 kHz. Field modulation sweeps through part of the ESR signal and reflected microwave is then amplitude modulated at the same frequency. The modulation amplitude should be much smaller than the width of the ESR signal, so that the signal is approximately linear over an interval of the same width as the modulation amplitude. The ESR signal is then transformed into sine wave with amplitude proportional to the slope of the signal as shown in Figure 23. The signal channel produces a DC signal proportional to the amplitude of the modulated ESR signal. Then it compares the modulated ESR signal to the reference signal and suppresses everything else that has a different frequency or phase than the field modulation. To improve sensitivity even more time 62 Figure 23 Field modulation and phase sensitive detection constant is used to filter out the noise. This way sensitivity can be increased by several orders of magnitude, but modulation amplitude, frequency and time constant must be appropriate or the signal will be distorted. The magnetic field amplitude must be smaller than the width of the ESR signal or the signal will be broadened. The time constant filters out the noise by slowing the response of spectrometer. However, if the time constant is too long, the ESR signal itself can get filtered out. Also the magnetic field at which resonance occurs can become shifted as well, preventing us from finding the correct g value. Sometimes it is necessary to use a long time constant to detect a very weak signal. In this instance the sweep time must be about 10 times longer than the time constant. Distortion of the signal (broadening) can occur if the modulation frequency is too high. However, this is a only concern when dealing with dilute solutions and samples that have very narrow or closely spaced resonances. 2.3.3.4. Magnetic field controller The magnetic field controller consists of two parts. The first part sets the magnetic field value and timing of the field sweep. The second part regulates the current in magnet 63 windings to reach the requested magnetic field value. Magnetic field regulation occurs via a Hall probe placed in the gap of the magnet. It produces voltage that depends on magnetic field perpendicular to the probe. Regulation of the magnetic field is achieved by comparing the voltage from the Hall probe to the voltage given by the first part of controller 2.3.4. Light induced ESR (LESR) This study focuses on photogenerated species, so rather than studying ESR we study light induced ESR. The experimental set up is identical except for introduction of a laser. The experiment consists of two parts. First we measure the ESR before illumination of the sample with the laser light. Then we measure the ESR again while illuminating the sample with the laser. We call this measurement light on measurement. The final result or LESR spectrum is obtained by subtracting ESR from light-on measurement. 2.4. Optically detected magnetic resonance (ODMR) spectroscopy The LESR technique is very successful in studying the blends of polymers and fullerenes. However, the LESR signal in pristine π-conjugated polymers is very low because of the low density of excited states and the sensitivity of shotky barrier diode used to detect the reflected microwave. The optical transitions change with magnetic resonance, so rather than observing changes in reflected microwave that are in order of meV we can observe changes in transmission light intensity that are in order of eV. So using optically detected magnetic resonance (ODMR) we enhance spin detectability by orders of magnitude. 64 There are two main techniques: PLDMR- the detection of changes in PL under magnetic resonance and PADMR - the detection of changes in photoinduced absorption under magnetic resonance. 2.4.1. PADMR Excited state recombination is a spin dependant process. The resonant absorption of microwave induces a spin flip, redistributes excited state populations among magnetic sublevels changing the recombination rate. This results in a change of the optical properties, namely the absorption and the emission. PADMR is a desired technique when dealing with materials that have low photoluminescence, such as blends of polymers and fullerene derivatives. However, since PADMR is based on a change of the recombination rate we can only observe spin pairs and not individual spins as in the case of LESR. The photoinduced absorption spectrum shows the optical transitions and PADMR distinguishes if these transitions are due to spin ½ (polarons) or spin 1 (triplet exciton). For spin ½ species PADMR, the change of excited states, N due to magnetic resonance can be expressed as in the equation (2.78). (2.78) Subscripts P and AP mean parallel and antiparallel spin pairs; tilde above N, the number of spin carriers denotes the resonant condition. The number of excited states N can be expressed in terms of generation rate, G and recombination rate, R. 65 (2.79) Under microwave saturation the number of parallel and antiparallel spins is equal. (2.80) It follows that (2.81) In the case of geminate recombination where all of the spins are generated in antiparallel configuration NP=0 and NAP=N, equation 2.91 becomes (2.82) In the case of distant pair recombination where all of the spin ½ pairs are randomly generated, we can assume that GAP=GP and therefore (2.83) From here the equation (2.82) becomes 66 (2.84) From equation (2.84) we can conclude that the sign of PADMR can tell us more about the recombination of the spin species. However, the geminate pairs have too short a lifetime to significantly contribute to CW PADMR. For spin 1, triplet exciton, we have to consider spin-spin interaction. So the spin Hamiltonian can be expressed with equation (2.85). (2.85) where S is the sum of two interacting spins, D and E are zero-field splitting (ZFS) parameters derived from coupling of two spin ½ particles responsible for triplet energy splitting at zero magnetic field. The spin 1 PADMR is a consequence of microwave induced changes in population of the triplet energy sublevels. There are three transitions occurring, two full-field transitions for Δm=1 and one half-field transition corresponding to Δm=2 that is only possible because of the anisotropic spin exchange interaction that mixes the triplet sublevels as depicted in Figure 24.The magnetic resonance satisfies equation (2.86): (2.86) 67 Figure 24 Spin 1 resonance transitons where θ and ϕ are the Eulers angles for the principal axis of triplet‟s zero field splitting D tensor with respect to the magnetic field coming from spin-spin dipolar interaction. H0 is ћω/gβ. At zero magnetic field, the degenerate energy states corresponding to ms=1, ms=0, and ms=-1 are related to ZFS parameters in the following way: (2.87) (2.88) (2.89) 68 where zero field splitting parameters can be expressed by means of the triplet wavefunction, and single particle operators r12, x12, y12, and z12. (2.90) (2.91) In polycristaline and amorphous materials the D tensor is randomly oriented resulting in a PADMR spectrum called a triplet powder pattern that is the integration average over resonance conditions for all possible orientations. 2.4.2. PLDMR PLDMR is very successfully used in systems with strong photoluminescence because when the spin carriers recombine radiatively, the direct changes in PL can be observed. PLDMR for spin ½ can be expressed as: (2.92) where I is intensity of PL and ON and OFF subscripts denote microwave being on and off. (2.93) (2.94) 69 R is the recombination rate and R‟ is the radiative recombination rate. From here we can calculate that (2.95) PLDMR for spin 1, triplet excition is very similar to spin 1 PADMR. Its spectrum is a triplet powder pattern with half-field and full-field transitions and can be expressed as Δ Δ (2.96) 2.4.3. Experimental setup The ODMR experimental set-up is identical to PA experimental set-up with addition of superconducting magnetic and microwave as presented in Figure 25. The sample is mounted in 3GHz coaxial microwave cavity with windows for optical access that is placed between poles of the superconducting magnet. The structure of the microwave cavity is shown in Figure 26. The electromagnetic field inside of the cavity can be described by the Maxwell equations: (2.97) (2.98) 70 laser mirror mirror mirror mirror light source monochromator detector preamplifier lock in amplifier power supply magnet controller microwave modulator microwave source function generator sample cryostat magnet Figure 25 ODMR experimental set-up where (m=1,3,5,…), , r1 and r2 are the inner and outer radii of coaxial resonant cavity, P is microwave power, Q is quality factor of resonant cavity, L is the length of cavity, resonant frequency is . In the system used r1=0.627cm, r2=2.2cm, L=9.9cm, ωres=2P×2946MHz, and Q=1000. The maximum power of the microwave is 100mW which produces inside of the cavity electric field of 90V/cm and magnetic field of 0.3G. The temperature in the cavity is adjusted manually by changing the flow of liq. He into sample chamber or via heater placed on the sample holder. The superconducting magnet provides a slowly swept DC magnetic field reaching 3T (30,000 G). The magnet is powered by a HP6260B DC power supply controlled by AMI 400A 71 Figure 26 The 3 GHz coaxial resonant cavity programmer that is in turn controlled by Keithley 236, a stable voltage source with computer interface (GPIB). As the microwave source a HP616B UHF high frequency signal generator is used. Microwave is amplified by minicircuit ZHL-42 amplifier, modulated by a HP 11720A pulse modulator which is controlled by TTL signal from a function generator. Amplified microwave is transmitted through a Raytheon CSL71 bridge via a coaxial wave guide and it is coupled in resonant cavity via a copper antenna. The reflected microwave converted in voltage is monitored by an oscilloscope. The reflected signal is then compared to the ground signal to examine weather the optimum resonance conditions are achieved as shown in Figure 27. 72 Figure 27 Oscilloscope display - achieving optimum resonance conditions, the maximum reflected microwave signal 2.5. Polaron pair mixing under magnetic resonance In this study we pay special attention to the blends of polymers and fullerene derivatives where the dominant photoexcitations are due to polarons. It is important to discuss what happens to polaron pairs under the influence of magnetic field and microwave, as is the case in magnetic resonance experiments. In polaron pair triplet (PPT) manifold due to Zeeman splitting there are three spin sublevels corresponding to m=+1,0,-1that are in resonance with microwave energy, ћω. Consequently transitions between m=0 and m=±1are induced as shown in Figure 28. The triplet m=0 sublevel is coupled with singlet PPS spin sublevel via intersystem conversion interaction that is mainly determined by the hyperfine interaction. Through the resonant microwave all of the spin sublevels S, T-1, T0, and T+1 of PP are interconnected. At off resonance conditions triplet spin sublevels with m=±1 mix strongly with singlet spin sublevel. However, the population of triplet and singlet spin sublevels with m=0 is governed by their own generation and decay rates. 73 Figure 28 Transitions within polaron pair spin sublevels under the influence of a magnetic field. Microwave induced transitions under the magnetic field are depicted by the arrows between spin sublevels. KISC represents mixing of spin states within polaron pairs. CHAPTER 3 STUDY OF PHOTOEXCITATIONS IN DONOR MATERIALS 3.1. Poly(3-hexylthiophene) 3.1.1. Introduction Since the 70s, conjugated polymers have been extensively studied driven by possible applications in electronic, spintronics and optoelectronic devices. Polythiophenes form very environmentally and thermally stable materials, making them desirable for use in many applications such as donor materials in organic photovoltaic cells.[35] A very promising characteristic of the polythiophenes is that their structure and properties can be controlled by their synthesis. The polymer chain structure especially plays a critical role in determining the physical properties of this conducting polymer class. Their asymmetrical structure causes the polymerization to produce a mixture of three possible regio-chemical linkages between the repeat units: head-to-head, tail-to-tail (these are known as regio-random coupling), and head-to-tail or regioregular coupling. Regio-random couplings are considered to be less pure because they somewhat disrupt conjugation, and prevent ideal solid state packing, thereby diminishing the polymer electronic and photonic properties. On the other hand, regioregular coupling allows access to low energy planar conformation that results in highly conjugated polymers that provide flat, stacking macromolecular structures that can self-assemble into lamellas perpendicular to the substrate, and therefore providing efficient interchain and intrachain 75 conductivity pathways. In other words regioregularity maximizes the electronic and photonic properties of many conducting polymers and in particular the polythiophenes [36]. Regioregular poly-(3-hexylthiophene) or RR P3HT has been successfully used in organic photovoltaic cells as a donor material. It has been demonstrated that the degree of regioregularity [37], as well as the architecture of an organic solar cell [38] affect the solar power conversion efficiency. In this study we concentrate on the effect of regioregularity on the photophysics and electrical properties of P3HT. It has already been shown that regioregularity increases with molecular weight and that charge transport amplifies with regioregularity (MW). [39],[40],[41],[42],[43] We will show that the photoinduced absorption spectrum can be used to differentiate between low and high molecular weight RR P3HT polymer chains, by tracking the appearance of a polaron band at low photon energy. Photo induced detected magnetic resonance (PADMR) is used as a complementary technique to understand the spin nature of the various PA bands in the spectrum. PADMR shows that the PA band in RRa P3HT is of excitonic (spin 1) nature. In the highest molecular weight RR P3HT the dominant PA band is due to delocalized triplet exciton. Moreover ultrafast transient PA shows that as the MW of P3HT chains decreases, a new PA band with different dynamics appears in the spectrum at 0.55 eV. Moreover, in high MW RR P3HT we also see stimulated emission and laser action that is absent in low MW polymer. Consequently it is possible by using these optical and magnetic probes to evaluate and predict the suitability of the P3HT polymer batches for their use in photovoltaic cells. 76 3.1.1. Study of absorption and emission spectra In Figure 29 the absorption and photoluminescence (PL) spectra of RR P3HT and RRa P3HT are shown. The absorption band over the optical gap is due to π-π* transitions. According to the normalized absorption spectra of the superior high molecular weight RR P3HT (50kDa) and RRa P3HT it is noticeable that RR P3HT has stronger absorption in the near IR and visible spectral region. According to Kasha‟s rule the PL emission comes from the lowest exciton in the system. In agreement with previous studies [44] we detect red shift of both absorption and PL of RR P3HT comparing to RRa P3HT, that may be assigned to the better order in lamellae and longer conjugation lengths [45]. The absorption and PL spectra of RR P3HT show pronounced structures due to phonon replica suggesting more homogeneity of polymer chains in the film[44]. The ratio of 0-0 to 0-1 vibronic side band in the PL spectra of RR P3HT is smaller than in RRa P3HT, suggesting a higher degree of aggregation of the former polymer.[46, 47] The PL intensity of RR P3HT is much lower compared to that of RRa P3HT. This is also reflected in the PL quantum efficiency (PLQE), which we measured to be approximately 4% for high molecular weight RR P3HT and 13% for RRa P3HT. This could be explained by a weaker radiative transition of the lowest lying excitons that may be due to a larger interchain contribution compared to intrachain excitons in RRa P3HT. On the other hand, comparing the latest high molecular weight RR P3HT to RR P3HT from 2001/2002 that had PL QE less than 0.5% [44, 48] we may conclude that our new RR P3HT polymer might have fewer defects, and therefore yielding higher PLQE. 77 Figure 29 Normalized absorption and PL spectruas of RRa P3HT and RR P3HT thin films. The PL was measured at 80 K whereas the absorption was obtained at 300 K. 3.1.2. Studies of photoinduced and doping induced absorption spectra In order to understand the excitonic nature in the polymer, we studied ps photo modulation spectra of thin films of RR P3HT from ADS, and from Plextronics, Inc. with different molecular weights, as well as RRa P3HT, as shown in Figure 30. These studies were carried out by Mr. Bill Pandit in our group. In RR P3HT film we can notice a blue shift of the excitonic PA band at ~1eV (referred to as PA1). However, with lowering molecular weight (to ~15 kDa) we observed the formation of a new PA band at ~0.6 eV that we dub PA*; this PA band is very pronounced and red shifted to 0.55 eV in the ADS RR P3HT film. When comparing the dynamics of the two PA bands, namely PA1 and PA* (Figure 30 inset (a)), it is obvious that they are not exactly the same. The decay of PA1 is faster than that of PA* indicating that they belong to different photogenerated 1.0 1.5 2.0 2.5 3.0 3.5 0 5 10 0.0 0.5 1.0 RR-a P3HT RR P3HT RRa P3HT OD (normalized) PL (normalized) Photon Energy (eV) RR P3HT  PL 78 Figure 30 Transient PA spectra of a) RR P3HT with three different molecular weights; b) RRa P3HT. The insets show time decay at 0.95eV and 0.55eV, respectively. 0.2 0.4 0.6 0.8 1.0 0 1 2 3 0 2 4 6 8 RR-a P3HT PA1 104 (-) Photon Energy (eV) 0 50 100 150 200 250 0.0 0.5 1.0 0.95eV 0.55eV -Normalized Delay (ps) 104 (-T/T) High MW Low MW x 1.34 ADS x1.99 PA1 PA* RR P3HT 0 100 200 300 400 0.0 0.5 1.0 b) PA* PA1 -T Normalized Delay (ps) a) .. ':.'. ... .... j i\. != ..... :~::-....... . ................ .-0.".! .• :: .................... .. Ii . '~ . \. • • • ! •• • i ••• / I I ......... -~ ... -/' Lta 79 species. The transient PA of RRa P3HT shows two PA bands at 0.95 eV and 0.55 eV, respectively, that also do not share the exact dynamics. Similarly to RR P3HT, the PA band at 0.95 eV has a faster decay than that of the 0.55 eV band. We tentatively assign the PA band at 0.55 eV to instantaneously formed polarons since it is longer lived than the excitonic PA band at ~1 eV. [49] In Figure 31(a) we show the photoinduced absorption spectrum of RR P3HT film at 80 K. It is seen that the high molecular weight (50kDa) RR P3HT film does not have a PA band at low photon energy (assigned as P1) as observed in the literature [44, 48, 50]. In contrast the PA spectrum is dominated by a PA band at ~1.08 eV with a shoulder at about 1.2eV. This PA band has been previously assigned as P2, related to a photogenerated polaron species. However, the DIA spectrum of high MW RR P3HT presented in Figure 32 does not show any bands at this photon energy. The polaron DIA bands occur at ~0.2 eV and ~1.8 eV, respectively. From here we can infer that the PA band at 1.08 eV is not due to charge excitation, but may be due to delocalized triplet excitons. The PA spectrum of lower MW RR P3HT (that yields lower power conversion efficiency) shows the P1 band at 0.2 eV, somewhat smaller PA band at 1.08 eV and stronger shoulder at 1.2 eV that may be assigned to localized triplet exciton. ADS RR P3HT spectrum has even stronger P1 band at low photon energy and the localized triplet exciton band dominates the PA spectrum at 1.2 eV. The PA spectrum of RRa P3HT is dominated by only one PA band at around 1.36 eV that is assigned to be due to localized triplet excitons, which is in agreement to DIA spectrum presented in Figure 33, where we do not see any clear polaron bands at this photon energy. 80 Figure 31 CW PA spectra of a) RR P3HT with three different MW; b) RRa P3HT. The PA bands P1 and P2 of polarons, PP for polaron pairs, and T for triplets are assigned. 81 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 0.0 0.2 0.4 0.6 0.8 - photon energy (eV) P1 P2 Figure 32 DIA spectrum of RR P3HT thin film doped with I2 for ~20 s. 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 0.0 0.2 0.4 0.6 0.8 - photon energy (eV) RRa P3HT thin film doped with I2 Figure 33 DIA spectrum of RRa P3HT thin film doped with I2. 82 3.1.3. PADMR spectroscopy The PADMR of RR P3HT film measured at ћω=1.08 eV (Figure 34(a)) shows a 9 Gauss narrow asymmetric band at H0=1000 Gauss, corresponding to g=2, with shoulders similar to triplet powder pattern.[51] The inset of Figure 34 (a) presents the closer look at the resonance at 1000 Gauss. Even though the resonance is very narrow it does not have a typical shape of a spin -½ resonance. This can be explained by delocalized triplet excitons. According to equations 2.90 and 2.91 with increasing distance, r between the electron and hole that form the delocalized triplet exciton, its zero field splitting parameters, D and E decrease. For very small D and E the zero field splitting between the spin sublevels can be negligible due to the weak magnetic dipole interaction between the spin ½ carriers comprising the triplets. Therefore the two full field transitions occur at very close magnetic field values, as depicted in Figure 35 that results in a narrow, asymmetric resonance. Based on the broader range of PADMR that does not show the typical „half field‟ triplet resonance, we may question that the PA band in high molecular weight RR P3HT is due to the triplet excitons. The lack of „half field‟ transition can be explained by very small values of zero field splitting parameter D and E, which based on the equation (3.1) (the FF resonance steps) and the equation (3.2) (the FF resonance shoulders) were calculated to be 12 G and 6 G, respectively. (3.1) (3.2) 83 0 200 400 600 800 1000 1200 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 0 400 800 1200 1600 -0.5 0.0 0.5 1.0 1.5 2.0 PADMR (10 3 n/n) 940 960 980 1000 1020 1040 1060 0.0 0.5 1.0 PADMR averaged (103 n/n) Magnetic Field (Gauss) b) S=1 FF PADMR (10 3 n/n) Magnetic Field (Gauss) S=1 HF a) Figure 34 PADMR of: a) RR P3HT film showing delocalized triplet exciton resonance; and b) RRa P3HT with spin -1triplet powder pattern. FF and HF powder patterns are assigned. The inset in a) shows a closer look at spin-1 resonance due to delocalized triplet exciton with small zero field splitting parameters D and E. 84 Figure 35 Resonant transitions for delocalized triplet exciton with very small zero field splitting parameters D and E. In contrast, the PADMR of RRa P3HT at ћω=1.36 eV consists of a weak 13 G narrow spin 1/2 resonance at H0=1002 G, and characteristic spin 1 triplet powder pattern with strong „alf field‟ resonance H1=368 G and „full-field‟ resonance powder patter between 800 and 1200 Gauss. The anisotropic spin exchange mixes the triplet sublevels, permitting both Δms = ± 1 and Δms = ± 2 transitions at 3 GHz. That is the reason that both full field (FF) and half field (HF) powder pattern resonances are observed. We can calculate the zero filed splitting parameters from the HF and FF powder pattern singularities. (3.3) (3.4) 85 where D and E are zero field splitting parameters. Using the equations (3.3) and (3.4) we calculate D=607 G and E=70 G. 3.1.4. Conclusion Based on our results, the power conversion efficiency of organic solar cell does not depend on the architecture of a device alone, but also on the synthesis of the organic materials used for the active blend. In the past 10 years the RR P3HT regioregularity continuously improved, yielding higher power conversion efficiencies. PLQE of the newest RR P3HT increased almost 10 times, which can be attributed to fewer defects. The PA spectrum of high MW (superior regioregularity) RR P3HT does not show a PA band at low photon energy, previously assigned to P1 of separated polarons. The PADMR and DIA show that the dominant PA band in this polymer is due to delocalized triplet excitons. Ultrafast photomodulation shows that with decreased MW a new band at 0.55 eV appears in the PA spectrum that does not share the same dynamics as the band at 0.95 eV that belongs to excitons. We thus conclude that this PA band is either due to trapped excitons in defects, or to excitons in the polymer aggregates. 3.2. TAES-V polymer 3.2.1. Introduction The bulk heterojunction organic photovoltaic cells (OPVC) based on RR P3HT/PCBM blend have been reported with power conversion efficiency (PCE) around 4%. [38] However, it seems that with this particular donor/acceptor blend the PCE limits have been reached. There are many factors that limit performance of OPVC; one of them is absorption of the polymer donor. It seems that narrow absorption of RR P3HT in the 86 spectral range of 300-650 nm is the main factor that hinders further improvement of OPVC based on RR P3HT.[52] In general, the performance of OPVC is directly proportional to three parameters: open-circuit voltage (Voc), short -circuit current density (Jsc), and fill-factor (FF). Jsc depends on the absorbance of the donor material; thus so to increase Jsc low -band gap polymers have been successfully synthesized.[53-56] To achieve high PCE, in addition to the low band gap (<1.6 eV), the ideal donor material need also to show have high hole mobility and well-matched HOMO/LUMO levels with the acceptor material [57] in order to enhance Voc. For this reason, building blocks with lower-lying HOMO are introduced in the donor polymers. TAES-V is a novel low-band gap polymer composed of three co-polymers of the structure D-A-D (where D stands for donor and A for acceptor) synthesized by Plextronics, Inc, which when blended with PC70BM in OPVC gives record power conversion efficiency of 7%. In this study we used optical and magnetic resonance spectroscopy in order to better understand the nature of photoexcitations in this new polymer. 3.2.2. Optical study of TEAS-V copolymer As shown in Figure 36 the absorption onset of TAES-V is at ~1.6 eV (the optical gap is~1.75 eV) indicating the existence of some low band gap building blocks in this polymer. The absorption spectrum peaks at approximately 2.08 eV, which is closer to peak of photon flux irradiation at ~1.82 eV [56] than that of RR P3HT (having absorption peak at ~2.4 eV). Comparing to RR P3HT, the absorption spectrum of TAES-V is dramatically red shifted by ~0.3 eV; this shift improves sun light absorption of this 87 1.0 1.5 2.0 2.5 3.0 3.5 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 OD (normalized) photon energy (eV) TAES-V RR P3HT Figure 36 Normalized optical density (absorbance) spectra of TAES-V and RR P3HT. copolymer and consequently Jsc also increases, resulting in better PCE of OPVC based on this material donor. PL measurements were done with the pump excitation wavelength of 514 nm and 100mW power at 50 K (Figure 37). The PL spectrum is very well structured suggesting long conjugation lengths and higher regioregularity with the primary exciton emission (or 0-0 emission line) at ~1.75 eV. However, the PL spectrum shows very strong and almost equal in intensity 0-0 and 0-1 vibronic transitions; this indicates lack of polymer chain 88 Figure 37 Normalized PL and absorption spectra of TAES-V thin films. Inset shows the vibronic energy level diagram with the optical transitions corresponding to absorption peaks (A1,A2) and PL peaks (0-0, 0-1,0-2) . 89 aggregation, and contribution of intrachain excitons to PL. The ratio of 0-0 to 0-1 transition, Sr is close to unity, almost its maximum value; that is characteristic of single molecule emission. Consequently, the PL spectrum could be described with Franck Condon model giving relative intensity of vibronic replica [46]. (3.5) where nf is the real part of the refractive index that is usually taken as constant over the spectral range, m is the index of the vibrational level, and S is the Huang-Rhys factor, which gives a measure of the coupling between the electronic transition and the most strongly coupled vibration. The PL spectrum of TAES-V was fit with equation (3.5) using Gaussian line shape with the same width for each vibronic transition resulting in S=0.98±0.06 as shown in Figure 38) [46, 47, 58]. The photoluminescence quantum efficiency of this polymer films is relatively high ~17% compared to 4% of the best (highest molecular weight ) RR P3HT . This gives further evidence for the intrachain exciton recombination and the lack of aggregation. Photoinduced absorption (Figure 39) was measured on a drop-casted film of TEAS-V with concentration of 3 mg/ml in chloroform. Sapphire was used as a substrate. The pump wavelength was 514 nm with100 mW power. A lamp was used as a probe light. The photoinduced absorption spectrum shows no band due to polaron P1 at the low photon energy suggesting fewer defects. At ~1.06 eV there is a broad PA band and photobleaching starting at ~1.8 eV due to depletion of ground state population due to the 90 Figure 38 The fit of the PL spectrum using the Franck-Condon model (the equation (3.5)). The empty rhombs represent experimental data, and the red line is the fit. 91 Figure 39 Photoinduced absorption spectrum of TAES-V thin film with only one PA band at ~1.06 eV and photobleaching starting at 1.8eV. photogeneration of long-lived excitations. In order to distinguish the nature of PA peak at 1.06 eV additional experiments involving magnetic resonance need be done. Frequency dependent PA was done by changing the pump beam modulation frequency from 10 Hz to 100 kHz at a fix photon energy of 1.06 eV, and at two different pump powers, namely 100 mW (Figure 40) and 400 mW (Figure 41). The measured photoexcitation lifetime, 0.14 ms ± 0.05 ms does not significantly change with the pump power suggesting monomolecular recombination kinetics, as explained in chapter 2. Using Equation(2.30) a dispersive parameter, α may be obtained as~0.89. The PA intensity dependence is linear (Figure 42) which further proves monomolecular recombination kinetics. The decrease of PA intensity with increasing temperature (Figure 43) is related with a parallel decrease in the photoexcitation mean lifetime with the temperature. 92 Figure 40 Frequency dependent PA measured at a fixed photon energy of 1.06 eV, with laser power of 100 mW at 50 K. Figure 41 Frequency dependent PA measured at fixed probe photon energy1.06 eV with laser power of 400 mW at 50K. 0 5 10 15 20 25 30 35 40 1.00E+00 1.00E+01 1.00E+02 1.00E+03 1.00E+04 1.00E+05 in phase real out phase real in phase fit out phase fit 0 5 10 15 20 25 30 1.00E+00 1.00E+01 1.00E+02 1.00E+03 1.00E+04 1.00E+05 in phase real out phase real in phase fit 93 Figure 42 PA of TAES-V at 1.06 eV vs. the excitation laser power. The line through the data is a linear fit. Figure 43 PA of TAES-V vs. the temperature; the line through the data points is a linear fit to guide the eye. 94 From electroabsorption spectrum shown in Figure 44 we can identify the exciton at ~1.93 eV, lowest exciton state 1Bu. At ~2.45 eV we can recognize an induced absorption feature, assigned mAg. According to the separation of EA exciton contribution (1Bu) and continuum band (mAg) we can conclude that the exciton binding energy is ~ 0.5 eV. However, this spectrum does not show the signature of charge transfer (CT) exciton. Figure 44 Electroabsorption of TAES-V film (black line) follows the derivative of absorption (red line). 95 3.2.3. Magnetic resonance study of TAES-V copolymer Optically detected magnetic resonance (ODMR) in Figure 45 was measured at 10 K, with microwave power of 100 mW, laser wavelength 514 nm and power of 600 mW. The PLDMR was obtained with very good signal to noise ratio, which was not possible with PADMR measurement. The PLDMR measurement was done at fixed photon energy 1.06 eV, where we observed a broad PA band. PLDMR consists of a 9 G narrow spin-1/2 resonance at 1011 G and a spin-1 triplet powder pattern with over 500 G wide full field and ~20 G wide half field resonances. The H-PLDMR of TAES-V is reminiscent of PLDMR of an anisotropic exchange coupled polaron pair with the Hamiltonian expressed in equation (3.6)[59]: (3.6) where S and ms are the spin quantum numbers and . Equation (3.6) has one singlet or even solution and three odd or triplet solutions. HSS can be written in terms of zero-field splitting (ZFS) parameters, D‟ and E‟, where Triplet Hamiltonian can be expressed as: (3.7) The anisotropic spin exchange interaction mixes the triplet sublevels therefore permitting transitions corresponding to Δms=±1 and Δms=±2 as shown in Figure 46. 96 -200 0 200 400 600 800 1000 1200 1400 -1 0 1 2 3 4 106 x PLDMR H field (Gauss) TAES-V PLDMR laser: 514nm, 600mW doublet spin 1/2 triplet spin 1 full field triplet half field Figure 45 H-PLDMR spectrum of TAES showing spin1/2 resonance at 1011 G, and FF and HF triplet spin-1 powder patterns, as indicated. Figure 46 Schematic diagram of polaron pair energy levels with anisotropic exchange interactions. 97 Consequently, exchange triplet powder pattern has both full field (FF) and half field (HF) resonances. The following equation can give us singularities, shoulder, and steps of FF powder pattern: singularities: (3.8) shoulder: (3.9) steps: (3.10) The HF powder pattern consists of a singularity and shoulder that can be calculated as well. singularity: (3.11) shoulder: (3.12) Using equations (3.11) and (3.12) we calculate ZFS parameters D≈635G and E≈62G. A closer look at the spin 1/2 resonance reveals that it is asymmetric consisting of at least two contributions, possibly holes and electrons on the polymer chain, which confirms that an electron-withdrawing moiety has been added to this polymer backbone in order to lower HOMO level (Figures 47 and 48). PLDMR at various MW power was measured for both the spin-½ and HF spin-1 resonance (Figure 49), and the area under the resonance peak was plotted vs. MW power. The saturation behavior of the two resonances is different; the spin 1 HF 98 980 990 1000 1010 1020 1030 1040 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 TAES-V PLDMR spin 1/2 laser: 514nm, 400mW 106 x PLDMR H field (Gauss) 1011G Figure 47 PLDMR of TAES-Vfilm showing asymmetric spin-1/2 resonance. 980 990 1000 1010 1020 1030 1040 -1 0 1 2 3 106 x PLDMR H field (Gauss) Figure 48 Fitting spin ½ resonance with three Lorentzians indicating positive and negative contributions to the resonance 99 340 360 380 400 420 440 460 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 TAES-V PLDMR peak at HF triplet laser: 514nm, 400mW 106 x PLDMR H field (Gauss) Figure 49 HF resonance of TAES-V film due to Δm=±2 transition resonance saturates faster than the spin ½, which indicates two different photoexcitation species having different spin relaxation lifetimes (Figure 50). The data were fit with the saturation equation: , where Ps indicates the spin relaxation rate. And according to our results the Ps for spin ½ resonance is ~14, and for spin 1 HF resonance Ps is ~2. Since the spin relaxation time is reciprocally proportional to spin elaxation rate, we can conclude that triplet has much longer spin relaxation time than the polarons. Spin- ½ resonance is best fit with three Lorentzian curves as shown in Figure 48. We see two strong positive contributions and one negative contribution to the resonance. Since the ~3 GHz microwave frequency of our system does not allow us to clearly resolve these three contributions to the sharp spin ½ resonance, then LESR with microwave frequency of ~9 GHz proves to be a useful technique that also allows us to find the exact values of g-factors. 100 Figure 50 Integrated PLDMR resonance vs. MW power for spin1 and spin1/2 resonances. The lines through the data points are fits using the saturation equation: LESR of TAES-V film was measured at 10 K, using the Bruker Elyxis 580 X-band spectrometer. Ar+ ion laser was used as a pump with excitation wavelength of 514 nm and ~200 mW power. Microwave in this experimental apparatus is approximately ~9.67 GHz with power of 0.39 mW. First, we measured ESR at 10K. After that we directed the laser light onto the sample using fiber optics; we dub this measurement as „Light On‟. Intensity of ESR is very large, ~50% of the intensity of Light On measurement, as can be seen in Figure 51. The ESR signal is broader than Light On signal confirming a negative contribution to the resonance. Light Off measurement was done in the dark after the sample was already exposed to the excitation with the laser, 101 3430 3440 3450 3460 3470 3480 -3x105 -2x105 -1x105 0 1x105 H Field (Gauss) Figure 51 ESR, „Light On‟, and „Light Off‟ measurements of TAES-V film at 10 K using ~9 GHz microwave frequency and the laser was turned off. The long-lived LESR is slightly larger than the ESR signal, indicating the existence of metastable states. LESR is obtained by subtracting ESR from Light On measurement, showing only the contribution of photogenerated specie (Figure 52). Integrating LESR as shown in Figure 53 we obtain a graph similar to ODMR that has well resolved resonances. Transforming x axis from magnetic field into g-factor values we obtained the following g-factor for each contribution: 2.00078 ± 0.00008, 2.00214±0.00002, and 2.00537±0.00004. This measurement was done before we realized that we can calibrate system with DPPH to get exact g-factors. So the g-factors might be as well 1.998, 1.999, and 2.00237. The origin of negative contribution with g-factor of 2.00214 seen in both PLDMR and PADMR can be due to spin-dependant nonradiative trapping of photogenerated polarons by negative radical ions on acceptor moiety that might be responsible for large ESR signal[51, 60] . 102 3430 3440 3450 3460 3470 3480 -1.5x105 -1.0x105 -5.0x104 0.0 5.0x104 1.0x105 LESR H Field (Gauss) LESR TAES-V Figure 52 LESR of TAES-V film at 10K, obtained by subtracting the Light On ESR from the ESR in the dark . 1.990 1.995 2.000 2.005 2.010 2.015 -100 -50 0 50 100 integrated LESR g-factor Figure 53 Integrated LESR fit with three Lorentzian curves, showing two positive and one negative contribution with different g-factors. 103 3.2.4. Conclusion TAES-V is a novel low band gap (~1.7 eV) polymer composed of three copolymers of the structure D-A-D. When used in OPVC it gives 7% power conversion efficiency that can be partially due to its red shifted absorption. PL of this polymer is very well structured indicating long conjugated chains and high regioregularity. The 0-0 vibronic transition is very strong, almost equal to the 0-1 transition that is typical of single molecules, suggesting low aggregation of the undoped pure polymer. PLQE is very high, 17%, which can be attributed to intrachain exciton recombination. Photoinduced absorption shows only one PA peak at ~1.06 eV that is due to triplets. PLDMR clearly shows spin-1 triplet powder pattern with both full field and half field resonances. At 1011 Gauss there is another sharp asymmetric resonance due to spin-1/2, but we do not see it in PA. The PLDMR can be the result of both triplet and polaron pair resonances, where shoulders of polaron-pair might be buried in triplet powder pattern, giving us very large D and E zero field splitting parameters. Different saturation behavior of PLDMR with increase microwave power of half field resonance as opposed to the resonance at 1011 G confirms that they are due to different photogenerated species. Under closer inspection of the narrow spin-1/2 resonance we see that it consists of two positive and one negative contribution. Due to the small microwave frequency of our apparatus these contributions cannot be clearly resolved using PLDMR. When using LESR we are able to clearly resolve these contributions and confirm the negative resonance. The negative resonance may be due to nonradiative trapping of polarons by radical ions on the acceptor copolymer, which can be the cause of a large ESR signal. CHAPTER 4 ACCEPTOR MATERIALS IN ORGANIC PHOTOVOLTAICS 4.1. Introduction Some of the main requirements for good electron acceptor materials used in bulk heterojuction organic solar cells are large charge carrier mobility, large LUMO energy, and fast interfacial charge separation rates. Fullerene based materials have been thoroughly investigated since their discovery in 1985. The use of C60 as an acceptor in organic solar cell was first reported in 1993 by Sacriftci et al.[5, 61] Fullarenes have many attractive characteristics for OPV, such as a good electron transport and LUMO energy level that is both high enough to support a large Voc but also low enough to provide ohmic contacts for electron extraction and injection from common cathodes. Consequently, they became the acceptor materials of choice for the best performing organic solar cells.[2, 62, 63] C60 exhibits all the desirable characteristics mentioned above, but it is insoluble in common organic solvents. In order to improve its solubility and tune its electronic properties, functional groups such as butyric acid methyl ester are added to its cage structure. Each functional group attached lowers the electron affinity or increases the LUMO energy level of the acceptor by ~100 meV. Like C60, fullerene derivatives tend to aggregate and thereby assist phase segregation in blends with donor materials that is essential for charge separation 105 in OPVC. In this chapter our studies of PCBM, bis-[60]PCBM, and indene-C60 bisadduct will be presented. 4.2. PCBM 4.2.1. Introduction In an organic photovoltaic device excitons (usually Frenkel excitons) are formed in the donor materials following light absorption[18]. Exciton dissociation at the donor/acceptor interface is the critical step in harvesting electricity from light [2, 18, 38, 64]. However, it has been shown in the literature that charge separation in an organic material can be achieved without the donor/acceptor interface [18, 65]. Consequently, it is possible to produce OPVC with an active material consisting of PCBM only [66] with the power conversion efficiency up to 2.64%. [67] This is possible if PCBM itself shows phase separation that causes exciton dissociation, or excitons in PCBM dissociate at the interface with the electrodes or the hole transport layer (PEDOT-PSS). In this section we study aggregation of PCBM using optical and magnetic resonance studies of PCBM films and PCBM isolated in polystyrene matrix. 4.2.2. Sample preparation In order to make a PCBM film 18 mg of PCBM (supplied by Plextronics, Inc.) was dissolved in 1 ml of ortho-dichlorobenzene (ODCB) solvent. The PCBM dispersed in polystyrene was prepared in the following way: 100 mg of polystyrene was dissolved in 10 ml of toluene; and 10 mg of PCBM was dissolved in 10 ml of toluene. Subsequently the polystyrene solution was mixed with the PCBM solution achieving a ratio of 100:1. Samples were prepared by drop casting of the proper 106 solution on a sapphire substrate for PA measurement, and on a glass substrate for ODMR measurement. The annealing of chosen samples was done in inert nitrogen atmosphere at 150ºC for 30 min. 4.2.3. Optical studies of PCBM film and isolated molecules PCBM absorbs mainly in the UV region, as shown in Figure 54 resulting in a limited role in light harvesting in an OPVC. The absorption spectrum is mostly featureless. 1.4 1.6 1.8 2.0 2.2 0 1 2 3 4 5 absorption photon energy (eV) PCBM Figure 54 Absorption spectrum of PCBM film at room temperature, and PCBM molecular structure in the inset 107 Figure 55 shows the PL spectrum of PCBM dispersed in polystyrene. Since the ratio between PCBM and polystyrene is 1:100, this sample can be treated as containing isolated molecules. The PL spectrum confirms this, because the 0-0 vibronic side band is very strong suggesting the lack of aggregation [46, 47]. At ~2.0 eV we see an emission that can be attributed to the interaction of PCBM molecules and polystyrene, as it will be discussed later in this chapter. Consequently, we do not see this emission band in the PCBM film. Figure 56 shows the PL spectrum of the annealed sample of PCBM isolated molecules. The 0-0 vibronic side band became significantly weaker, causing the ratio between 0-0 and 0-1 vibronic sidebands to decrease, which is an indicative of the formation of aggregates [47]. The emission at 2.0 eV decreases as well. The well structured PL spectrum of the PCBM film has a dominant 0-1 vibronic side band and almost completely diminished 0-0 band. This provides further evidence for the formation of aggregates, as depicted in Figure 57. Moreover, the PL spectrum of PCBM film has red shifted and has become narrower compared to isolated PCBM molecules as can be seen in Figure 58 , suggesting increased order [47]. In addition to PL, XRD experiment on annealed film of PCBM was done by Dr. Tho Nguyen. The result (Figure 59) further supports the formation of aggregates in the PCBM film. 108 1.2 1.4 1.6 1.8 2.0 2.2 0 1 2 3 4 5 6 7 105 x PL photon energy (eV) 0-0 0-1 0-2 0-3 Figure 55 PL spectrum of PCBM dispersed in polystyrene at 45K with assigned vibronic side bands. The ratio between PCBM and polystyrene is 1:100 and therefore PCBM can be treated as isolated molecules. The excitation wavelength was 488 nm and laser power was 100 mW. 1.2 1.4 1.6 1.8 2.0 2.2 0 2 4 6 8 PLx105 photon energy [eV] 0-0 0-1 0-2 0-3 Figure 56 PL spectrum of isolated PCBM molecules in polystyrene, at 45K with the excitation wavelength of 488 and laser power of 100 mW. The sample was annealed at 150ºC for 30 min. 109 1.2 1.4 1.6 1.8 2.0 2.2 2.4 0 1 2 3 4 104 x PL photon energy (eV) 0-0 0-1 0-2 0-3 Figure 57 The well resolved PL spectrum of PCBM film at 45K with the excitation wavelength of 457 nm and laser power of 80 mW. 1.2 1.4 1.6 1.8 2.0 2.2 2.4 0 1 2 3 4 5 PL normalized photon energy (eV) 0-0 0-1 0-2 0-3 Figure 58 PL spectra of PCBM isolated molecules (black line), annealed sample of isolated PCBM (red line), and PCBM film (black line). 110 0 20 40 0 600 1200 1800 8 10 12 500 1000 1500 intensity 2 Figure 59 The XRD of annealed film of PCBM supports formation of ~8.5 nm large crystals. Experiment and calculation were done by Dr. Tho Nguyen. Figure 60 depicts the PA spectrum of PCBM in both isolated and film forms. In isolated PCBM (Figure 60(a)) the spectrum is dominated by a PA band (T1) at ~ 1.7 eV. In the PCBM film, however, the T1 band at