Role of Bad Dihedral Angles: Methylfluorenes Act as Energy Barriers

Sep 18, 2014 - For a more comprehensive list of citations to this article, users are ... toward Far-Red/NIR Region Using Different π-Spacer and Accep...
0 downloads 0 Views 721KB Size
Article pubs.acs.org/JPCA

Role of Bad Dihedral Angles: Methylfluorenes Act as Energy Barriers for Excitons and Polarons of Oligofluorenes Tomoyasu Mani* and John R. Miller* Chemistry Department, Brookhaven National Laboratory, Upton, New York 11973-5000, United States S Supporting Information *

ABSTRACT: “Defects” are one of the main obstacles for the use of organic conjugated molecules in efficient organic photovoltaics, but the definite origins of these defects are elusive to experiments and even in concept. Bad dihedral angles in conjugated molecules produced by adjacent units are considered to act as defects for excitons and polarons, slowing down their transports. While such defects are discussed, their properties were not well-understood. As a model system for such defects, we synthesized oligofluorenes incorporating methylfluorene(s) that can create large dihedral angles between adjacent fluorenes due to steric hindrance, mimicking bad dihedral angles presumably produced in polyfluorenes. Experimental measurements find that singlet excitons are substantially more sensitive to such bad dihedral angles than triplet excitons or negative or positive polarons. The barrier heights for singlets are about three times higher than the barriers for electrons, holes, or triplets. For all four species, the large dihedrals act as energy barriers, not traps.



INTRODUCTION

defects are discussed in the literature, their properties were not well-studied. To test the hypothesis, we study fluorene derivatives. Oligofluorenes and polyfluorenes are widely studied for the applications in OPVs13 and organic light-emitting diodes (OLEDs).14 In oligo- and polyfluorenes, “wire-like” transfer of charges8,10 and transport of excitons may be observed.9,15 While our focus here is on fluorene derivatives, it is reasonable to think that bad dihedral angle defects are present in many other types of polymers whenever dihedral angles fluctuate, principally due to thermal activation. The most stable conformations of conjugated polyfluorenes are twisted (ϕ ∼ 40°) in the ground state, while they tend to form more planar conformations (ϕ < 20−30°) in the excited states16,17 and charged states,18,19 enabling bigger electronic couplings. The energy barriers associated with dihedral angles are steeper in excited states and charged states.16,19 If bad dihedral angles are produced, preventing polymers to take a preferable conformation for transport, they might have a big impact on their couplings and transport capabilities of excitons and polarons. Venkataraman et al. showed that single-molecule junction conductances are highly dependent on dihedral angles between biphenyl’s two aromatic rings: more twisted less conductance.20 In longer oligomers and polymers, it is experimentally difficult to study only the effects of bad dihedral angles on transports of excitons and polarons as these angles are susceptible to thermal fluctuations. Twisting of dihedral angles in the backbone of a conjugated molecule typically occurs on the picoseconds

Organic conjugated polymers, “molecular wire”, are appealing candidates for low-cost and flexible photovoltaics because of their ease of synthesis and the tunabilities of their electronic/ structural properties by chemical synthesis. The power conversion efficiencies of organic photovoltaics (OPV) have improved over the past two decades, now approaching 10%.1−3 They are still far from the theoretical limit of ∼21% for single bandgap cells and ∼28% for tandem solar cells.4 It is widely thought that “defects” (or disorders) are present in organic conjugated polymers and they negatively affect transport of excitons and polarons after photoexcitation, thus limiting the efficiencies of power conversion. However, the definite origins and effects of these defects are elusive. Defects may include misconnections, kinks or bending, and coiling in longer polymer chains.5−7 The way these defects manifest in polymers depends on the lengths of polymers.5 Defects may raise the energy of an exciton or polaron, serving as a barrier to slow down transport or alternatively a trap to stop transport. Existing data on exciton and charge transport to end-groups8,9 and microwave conductivity5,10−12 indicate barriers may be the more relevant type of defects. However, molecular origins of such defects are not well-understood, and further investigations are required. Among these possible defects, we are testing the hypothesis that bad dihedral angles (or torsion angles) between adjacent units in conjugated molecules can act as defects for excitons and polarons, affecting their transport in polymer chains. Here, we define bad dihedral angles as the angles which make poorer electronic couplings between the two units than those at relaxed states. They can be produced transiently or permanently in segments of longer polymer chains. While such © 2014 American Chemical Society

Received: August 8, 2014 Revised: September 9, 2014 Published: September 18, 2014 9451

dx.doi.org/10.1021/jp5080626 | J. Phys. Chem. A 2014, 118, 9451−9459

The Journal of Physical Chemistry A

Article

Scheme 1. Synthetic Scheme of F-FMe-F and F-FMe2-Fa

a

(i) Pd(OAc)2/PCy3/pivalic acid, (ii) 1-bromohexane/KOtBu, (iii) Br2/FeCl3, and (iv) 2-(9,9-dihexyl-9H-fluoren-2yl)-4,4,5,5-tetramethyl-1,3,2dioxaborolane/Pd(OAc)2/SPhos/K3PO4.

timescale.16,21,22 Transport of excitons and charges typically occur on faster timescales and may therefore occur at fixed nuclear conformations. Therefore, ensemble measurements only allow us to obtain average rates of exciton and polaron transport. In the present paper, we engineered bad dihedral angles defects by chemical synthesis to study the effects of such defects. As a model system, we synthesized the oligofluorenes incorporating methylfluorene(s) that can permanently create large dihedral angles with adjacent fluorene units due to steric hindrance, mimicking transient bad dihedral angles in polyfluorenes. We studied the energetics of relevant excitons (singlet and triplet) and polarons (anion and cation) by using an array of experimental (e.g., chemical reduction/oxidation and pulse radiolysis) and computational techniques. We determined that such bad dihedral angles raise the energies of relevant excitons and polarons, showing that they can act as energy barriers for their transports.

at 9-positions was performed under the basic condition in the presence of KOtBu, followed by the bromination at 2- and 7positions. The final products were obtained by the SuzukiMiyaura couplings with Palladium acetate and SPhos as supporting ligands.27,28 F329 and 2-(9,9-dihexyl-9H-fluoren2yl)-4,4,5,5-tetramethyl-1,3,2-dioxaborolane30 were synthesized as previously reported. The complete synthetic procedures and the characterizations of new compounds were described in the Supporting Information. Chemical Reduction with Sodium Biphenyl. Chemical reduction was performed with sodium biphenyl in THF, following the established procedure.18 The experiments were performed under an inert argon atmosphere. Chemical Oxidation with NOPF6 or Thianthrene PF6. Chemical oxidation was performed either with nitrosonium ion or thianthrene cation (Th+). NOPF6 was purchased from Alfa Aesar and used as received. Thianthrene PF6 (ThPF6) was prepared as described in the literature.31 Oligofluorenes were dissolved in 4 mL of CDCl3 (150−200 μM). NOPF6 or ThPF6 was dissolved in CD3NO2 (70−200 mM). Solutions of oligofluorenes were titrated with NOPF6 or THPF6 solution in 2.0 or 5.0 μL increments, and UV−vis-NIR spectra were recorded after each addition. All the solutions were prepared under an inert argon atmosphere. Due to some spectral overlap with oxidizing or reducing agents, the uncertainties in the extinction coefficients are ±15%. Pulse Radiolysis. Experiments to determine the redox potentials were performed at the laser-electron accelerator facility (LEAF) at Brookhaven National Laboratory.32 Redox potentials were determined by an equilibrium method using pulse radiolysis.18,33 The compounds were dissolved either in THF (for reduction) or 1,2-dichloroethane, DCE (for oxidation) with 0.1 M tetrabutylammonium tetrafluoroborate (Bu4NBF4). Solutions of electron or hole donors were titrated with solutions of respective acceptors. Kinetic traces of donors and acceptors were recorded with each addition of donor solutions. From at least five kinetic traces with different acceptor concentrations, equilibrium constants for bimolecular electron transfer between donors and acceptors were determined. Free energy changes (ΔG) were then determined



EXPERIMENTAL METHODS General Information. All solvents and reagents were obtained from standard commercial sources and used as received. Silica gel (Sigma-Aldrich, pore size 60 Å, 70−230 mesh) was used for column chromatography. 1H and 13C NMR spectra were recorded with a Bruker Avance III spectrometer operating at 400.16 and 100.62 MHz, respectively. UV−visNIR absorption spectra were recorded with a Cary 5 spectrophotometer (Varian). Luminescence spectra were recorded on a PTI Time Master Fluorimeter with 814 photomultiplier detection system (Photon Technology International). Fluorescence quantum yields were determined by comparison with 9,10-diphenyanthracene (DPA) in cyclohexane (Φ = 0.97).23,24 Synthesis. The complete synthetic scheme is outlined in Scheme 1. 9H-Fluorenes were synthesized by the Pd-catalyzed methods.25,26 While tandem reactions to fluorenes did not proceed to completion as previously pointed out,26 individual palladium(II) acetate Pd(OAc)2/tricyclohexylphosphine PCy3catalyzed Suzuki-Miyaura couplings and the ring-closures were successful with good-to-excellent yields. Alkylation of fluorenes 9452

dx.doi.org/10.1021/jp5080626 | J. Phys. Chem. A 2014, 118, 9451−9459

The Journal of Physical Chemistry A

Article

Synthesis. To make experimental tests, we synthesized the oligofluorene trimers incorporating methylfluorene(s) whose structures are shown in Chart 1 and determined energetics for

from equilibrium constants. The redox potentials are reported versus Fc+/0 in the presence of 0.1 M Bu4NBF4. Computations. Computations were carried out with Gaussian0934 using different exchange correlation approximations in density functional theory (DFT). All hexyl groups are replaced by H atoms. The geometry optimizations are performed without symmetry constraints, unless otherwise noted. Linear response time-dependent (TDDFT) calculations are then performed for low-lying excited states to determine transition energies. Calculations used the 6-31G(d) basis set and the PCM solvation in THF.35−37 Visualization was performed with GaussView 5.0.9.

Chart 1. Molecular Structures of F3, F-FMe-F, and F-FMe2-F



RESULTS Methylfluorenes as a Model System. We first calculated the energy barriers associated with rotation between two fluorene units in F2 and F-FMe2 dimers as simple model molecules. They are estimated by computation for the ground states at the level of 6-31G(d)/B3LYP38,39 (Figure 1). The

addition of positive or negative charges and singlet or triplet excitons to each oligomer. F-FMe-F and F-FMe2-F were obtained with the overall yields of 21 and 18% from the starting materials as outlined in Scheme 1. 1H and 13C NMR and elemental analysis confirmed the structures and purities of FFMe-F and F-FMe2-F. 2D NMR (COSY and TOCSY) showed the presence of methyl groups in the central fluorene units (see the Supporting Information). Although it is not a main point of the paper, it is worthwhile noting that the biaryl compounds 2′-bromo-2,5-dimethylbiphenyl and 2′-bromo-2,5,5′-trimethylbiphenyl form atropisomers,40 existing in the equilibrium between M and P, and giving rise to the complex 1H and 13C NMR spectra. DFT calculations estimate that M isomer is slightly more stable than P for the both compounds. While we did not determine the exact conformations, based on the integrals of 1H NMR spectra, we can estimate that diastereomeric ratios (dr) are 78:22 and 77:23 for the compounds 2′-bromo-2,5-dimethylbiphenyl and 2′-bromo-2,5,5′-trimethylbiphenyl, respectively. Excitons. The energy levels of excitons were determined by using absorption and emission spectroscopy. Absorption and fluorescence spectra in THF are shown in Figure 2a. The oligomers with methylfluorene(s) have the blue-shifted absorption bands of S0 → S1 transition with smaller extinction coefficients. The emission bands corresponding to the S1 → S0 transition are also blue-shifted for F-FMe-F and F-FMe2-F. F3 and F-FMe-F both exhibit vibrational spacings characteristic of the double-bond (CC) vibrations, while the emission spectrum of F-FMe2-F is less structured. Findings on vibrational spacing are summarized in Table S1 of the Supporting Information and further discussed in the later section. Quantum yields (Φ) are determined to be 0.96, 0.94, and 0.88 for F3, F-FMe-F, and F-FMe2-F. The energies of the first singlet excited states (0−0 transition) are obtained by the crossing points of absorption and emission spectra: 3.26, 3.40, and 3.48 eV for F3, F-FMe-F, and F-FMe2-F. To determine the energies of the lowest triplet excited states (T1), phosphorescence spectra were recorded in 2-MeTHF at 77 K and are shown in Figure S1 of the Supporting Information. The triplet energies are 2.28, 2.32, and 2.33 eV for F3, F-FMe-F, and FFMe2-F. Absorption and emission spectra of monomers were measured as well and shown in Figure S2 of the Supporting Information. The energies of the first excited state decrease in the order of FMe2 < FMe < F1. Polarons. The absorption spectra of anion, dianion, and cation are obtained by chemical reduction and oxidation (Figure 2, panels b and c). Although the transition energies of polarons are not direct measures of energy levels, they can

Figure 1. (a) Molecular structures of F2 and F-FMe2. (b) Relative energy vs the dihedral angle ϕ. Structures are optimized at each fixed dihedral angle. (c) Relative population vs the dihedral angle ϕ. Relative population was estimated by Boltzmann distribution for room temperature (295 K).

results show that methyl groups at 3,6-positions of fluorene ring produce larger dihedral angles than hydrogens. The energy minima are at 37° and 52° for F2 and F-FMe2, respectively. Judged from relative populations estimated by Boltzmann distributions, unlike normal fluorenes, F-FMe2 is very unlikely to take a planar conformation and they are rather prone to take more perpendicular conformations, suggesting that the introduction of methyl groups at the 3- or 6-position can create bad dihedral angles. For charged states (anion and cation), F-FMe2 is similarly computed to be more twisted than F2. 9453

dx.doi.org/10.1021/jp5080626 | J. Phys. Chem. A 2014, 118, 9451−9459

The Journal of Physical Chemistry A

Article

tabulated in Table 1, together with the calculated first transition energies of each species. The second transition energies (P2 bands) of anions and cations are reported in Table S2 of the Supporting Information and those calculated with B3LYP are reported in Table S3 of the Supporting Information. P1 transitions of anions and the first transition of dianion are lower in energy in F-FMe-F and F-FMe2-F, while P2 transitions are higher in energy in F-FMe-F and F-FMe2-F. The extinction coefficients of P2 transitions are slightly smaller as well. No clear shifts were observed for P1 transitions of cations while P2 bands of F-FMe2-F experienced considerable blue-shift compared to those of F3 and F-FMe-F. The values and trends of the transition energies determined experimentally agree well with the calculated results using the long-range corrected ωPBE functional41 (ω = 0.1 bohr−1) in THF as well as B3LYP. In general, ωPBE functional better followed the observed experimental results than B3LYP, which is consistent with our previous results on longer oligomers.18,42 The differences between experimental values and computed ones for the first transition energies are