Long-Lived Singlet and Triplet Charge Separated States in Small

Article pubs.acs.org/JPCC

Long-Lived Singlet and Triplet Charge Separated States in Small Cyclophane-Bridged Triarylamine−Naphthalene Diimide Dyads Conrad Kaiser, Alexander Schmiedel, Marco Holzapfel, and Christoph Lambert* Institut für Organische Chemie, Wilhelm-Conrad-Röntgen Research Center for Complex Material Systems, and Center for Nanosystems Chemistry, Universität Würzburg, Am Hubland, 97074 Würzburg, Germany S Supporting Information *

ABSTRACT: Two different dyads containing a triarylamine (TAA) donor and a naphthalene-1,8:4,5-bis(dicarboximide) (NDI) acceptor bridged by either a [2.2]- or a [3.3]paracyclophane (CP) were synthesized. These dyads show a high population of long-lived charge separated (CS) singlet and triplet states. The lifetimes of these different spin states vary only by 1 order of magnitude. This unique situation is a consequence of both a large electronic coupling V and a large exchange coupling 2J. The population of the different CS spin states and therefore the charge recombination (CR) and intersystem crossing (ISC) kinetics were monitored by standard ns-transient absorption spectroscopy. Together with fs-transient absorption spectroscopy supported by electrochemistry, steady state fluorescence and steady state absorption spectroscopy a detailed model of the photoinduced processes was derived.

■

INTRODUCTION

energy of ca. 0.5 eV would be enough to be useful for practical applications.1a,3a In this work, we employed triarylamines (TAA) as strong donors. They are well-known for their application in the XEROX process and their use as hole transport materials in other optoelectronic devices.5 Furthermore, if they are substituted in the para-position, they will produce very stable radical cations which generally show a characteristic strong and sharp absorption band in the 13 000−15 000 cm−1 (770−670 nm) region.6 In addition, their redox potential can be tuned by these substituents in the para-position.6,7 As strong acceptors we used naphthalene-1,8:4,5-bis(dicarboximide) (NDI). Some of its substituted derivatives are promising candidates as air stable semiconductors with high electron mobility for the application in organic field effect transistors.8 A particular property of NDI is to undergo fast ISC (ca. 10 ps in chloroform)9 without the use of heavy metals.10 However, this is only true for N-alkyl substituted NDI while in N-aryl substituted NDI the formation of CT states is faster (0.5 ps in chloroform) but can be followed by charge recombination to yield locally excited triplet states.9 In addition, NDI may form a stable radical anion which shows a sharp and intense absorption band at 21 000 cm−1 (475 nm) and a low intensity band at about 16 500 cm−1 (605 nm). These properties and in particular its steady state absorption can be electronically tuned by core substituents.8d,10b,11 Furthermore, due to the nodes of the π-HOMO and π-LUMO along the molecular long axis, substituents connected via the nitrogen atoms are electronically

One of the most important issues in the design of new organic materials for artificial photosynthesis, molecular electronics, or for photovoltaic or other optoelectronic applications is how to influence photoinduced charge separation, charge transport and recombination by using different bridges and the choice of different electron donors or acceptor1 The most promising approaches start with preparing systems as small and simple as possible and to try to understand them before moving on to bigger, more complex systems.2 In this study, we focus on small donor-bridge-acceptor dyads based on triarylamines, cyclophanes, and naphthalene diimides that shall produce long-lived charge separated (CS) states upon photoexcitation. The photoinduced ET dynamics in such systems is supposed to be a consequence of relative state energies, reorganization energies of the associated processes and electronic couplings between the different states.2a An additional though often neglected aspect concerns the spin multiplicities of the states involved as well as their spin interconversion.2d,3 An optimal combination of these approaches can ensure a high population (quantum yield) and a prolonged lifetime of the CS states. In order to achieve this, charge separation must be fast but charge recombination slow (ratio >104).3a,4 In particular, charge recombination can be slowed down by Marcus inverted region effects or by spin selection processes, small electronic coupling andin the inverted regionsmall reorganization energies.2a In contrast, fast charge separation can be enhanced by strong electronic coupling and small reorganization energies (as long as the charge separation process is in the Marcus normal region which is often the case). A CS state with a lifetime of some μs and an © 2012 American Chemical Society

Received: May 7, 2012 Revised: June 25, 2012 Published: June 26, 2012 15265

dx.doi.org/10.1021/jp304391x | J. Phys. Chem. C 2012, 116, 15265−15280

The Journal of Physical Chemistry C

Article

decoupled from the central π-system, therefore delaying charge separation and recombination processes.8b,11j,12 Another favorable side effect of this weak electronic interaction is that local electronic transitions and the reduction potentials of Nattached NDI are left unchanged when altering bridging molecules or the donor strength in dyads that contain NDIs.8b,c,10a,11a,h,i,13 In the dyads investigated in this work the above-mentioned donor and acceptor groups are bridged by cyclophanes (CP) because they are one of the smallest model systems for through-space interaction in π-stacks, an issue of great importance for organic electronic materials.14 Furthermore, cyclophanes are compact and rigid units that reduce the electronic coupling compared to fully conjugated bridges and thus are expected to slow down charge separation and recombination processes between the TAA and NDI even further.14g,15 Another important advantage resulting from our choice of redox chromophores is that the spectra of the main absorption bands of the TAA radical cation and NDI radical anion, which form if charge separation takes place, do not overlap with the ground state absorption of the dyads which ensures unequivocal characterization of the CS states. In this study we designed two different dyads A and B with the above outlined building blocks that only differ in the CP bridging unit, which is a [3.3]paracyclophane in A with a separation of the benzene rings of 3.3 Å16 (twice the van der Waals radius of carbon is 3.4 Å17) and a [2.2]paracyclophane in B with a distance of 3.1 Å18 between the benzene planes. The photophysical properties of these dyads will be compared with those of dyad C in which the TAA donor and the NDI acceptor are directly linked by a butadiyne unit as well as with those of the fragments D and E.

■

as the solvent. All potentials were measured against the redox couple of ferrocene/ferrocenium (Fc/Fc+). Spectroelectrochemistry. These measurements were performed in a cylindrical quartz cell with an optically flat cell bottom in which the three electrode setup consisted of a platinum disk working electrode (6 mm diameter), a goldcoated metal plate as counter electrode and an Ag/AgCl pseudoreference electrode. Measurements were performed under argon atmosphere with the same solvent and supporting electrolyte as in the cyclic voltammetry. The potentials were applied using an EG&G Princeton Applied Research Model 283 Potentiostat with potential steps of either 25, 100, or 250 mV. UV/vis/NIR spectra were recorded using a JASCO V-570 spectrometer in reflection at the polished working electrode which was adjusted to ca. 100 μm above the cell bottom by using a micrometer screw. Femtosecond-Transient Absorption Spectroscopy. The pump−probe experiments were performed in a 2 mm quartz cuvette equipped with a microstirrer. The target compounds were dissolved in Uvasol solvents from Merck and purged with argon for 30 min. The laser system consists of an ultrafast Ti:sapphire amplifier (Newport-Spectra-Physics, Solstice) with a central wavenumber of 12 500 cm−1 (800 nm), a pulse length of 100 fs and a repetition rate of 1 kHz. One part of the output beam was used to seed an optical parametric amplifier (Newport-Spectra-Physics, TOPAS) as the source for the pump pulse with an attenuated energy of 150−210 nJ, a wavenumber of 28 200 cm−1 (355 nm) and a pulse length of 140 fs. A small fraction of the Ti:sapphire output was focused into a moving calcium fluoride plate to produce a white light continuum between 12 500 cm−1 (800 nm) and 24 000 cm−1 (417 nm) which acted as the probe pulse. The depolarized excitation pulse was collimated to a spot which is at least 2

EXPERIMENTAL SECTION

Steady State Absorption Spectroscopy. These measurements were performed in 1 cm quartz cuvettes using a JASCO V-570 UV/vis/NIR spectrometer. The substances were dissolved in Uvasol solvents from Merck and the pure solvent was used as a reference. Absorption spectra were recorded in toluene within a concentration range from 1 × 10−5 to 3 × 10−5 M. Steady State Fluorescence Spectroscopy. These measurements were performed using a Photon Technology International (PTI) QM fluorescence spectrometer. Cuvettes (1 cm) were used and spectra recorded in Uvasol solvents from Merck after purging the samples for 30 min with argon. The lifetimes were measured using a PTI TM fluorescence lifetime spectrometer with a 340 nm laser diode for excitation. Colloidal silica in deionized water was used as a scatterer to determine the instrumental response function. Lifetimes were determined by fitting the decay curves (deconvolution with the instrument response function) with a single-exponential decay function. Cyclic Voltammetry. These measurements were performed using a BAS CV-50W electrochemical workstation in dichloromethane and tetrabutylammonium hexafluorophosphate (ca. 0.2 M.) as supporting electrolyte. In a threeelectrode setup, a glassy carbon electrode (3 mm diameter) working electrode, a helical platinum counter electrode and a platinum pseudoreference electrode were used in a sealed glass vessel flushed with argon. Dichloromethane, freshly distilled over calcium hydride onto activated aluminum oxide was used 15266

dx.doi.org/10.1021/jp304391x | J. Phys. Chem. C 2012, 116, 15265−15280

The Journal of Physical Chemistry C

Article

Scheme 1. Synthesis of [3.3]paracyclophane 10a

a Key: (a) Br2, Fe (1.8 mol %), 10−15°C → room temperature, 23 h; (b) NBS, AIBN, CCl4, reflux, 80 min; (c) KCN, EtOH/H2O 2:1, 70°C, 3.5 h; (d) H2SO4 (98% in water), MeOH, reflux, 2 d; (e) LiAlH4, THF, room temperature → 50 °C, 2 h; (f) HBr (33% in AcOH), 110°C, 16 h; (g) KSAc, THF, 40°C, 16 h; (h) KOH, MeOH/THF 1:1, reflux, 19 h; (i) H2O2 (35% in H2O), AcOH, 100°C, 16 h; (j) vacuum flash pyrolysis, 0.015 mbar, 550°C. Yields in brackets are estimated by NMR of the crude product and refer in case of 8 and 9 to not fully characterized isomer mixtures.

a sequential model (i.e., unbranched unidirectional model) modeling the IRF (ca. 200 fs), the white light dispersion (chirp), and the coherent artifact (the model used has the time characteristics of the IRF) at time zero to yield the evolution associated spectra (EAS) and their corresponding time constants.22 The number of components was estimated by singular value decomposition. Nanosecond−Transient Absorption Spectroscopy. The pump−probe experiments were performed in a 1 cm quartz cuvette. The target compounds were dissolved in Uvasol solvents from Merck and purged with argon for 30 min. The laser system consists of a Nd:YAG laser (Continuum, Minilite II) with a repetition rate of 10 Hz. As source for the pump pulse either the third harmonics of the 9400 cm−1 (1064 nm) fundamental 28 200 cm−1 (355 nm) was used directly or shifted by a Raman shifter (hydrogen, 50 cm path length, 50 bar) to 24 000 cm−1 (416 nm) and selected by a Pellin-Broca prism. The laser pulse energy was varied between 0.20 to 5.2 mJ with a pulse length of 5 ns. As the probe light source (xenon arc lamp), sample chamber and detector (photomultiplier tube) we used a modular system from Edinburgh Instruments (LP 920-K Laser Flash spectrometer). In general the recorded transient signal consists of transient absorption, fluorescence and ground state bleaching. The detected transient signal intensity (I(λ,τ)) with pump pulse and probe light is corrected by the fluorescence intensity (IF(λ,τ)) measured without the probe light. From the corrected transient signal intensity in combination with the intensity (I0(λ)) without the pump pulse the optical density can be recorded as ΔOD = −log[(I(λ,τ) − IF(λ,τ))/(I0(λ))]. The instrument response function of the setup was determined with an empty cuvette.

times larger than the diameter of the spatially overlapping linearly polarized probe pulse. After passing the sample the probe pulses were detected via a transient absorption spectrometer with a CMOS sensor (Ultrafast Systems, Helios). Part of the probe light pulse was used to correct for intensity fluctuations of the white light continuum. A mechanical chopper, working at 500 Hz, blocks every second pump pulse, in order to measure I and I0. By comparing the transmitted spectral intensity of consecutive pulses [I(λ,τ),I0(λ)] the photoinduced change in the optical density can be recorded as ΔOD = −log[(I(λ,τ)/(I0(λ))]. The relative temporal delay between pump and probe pulses was varied over a maximum range of 8 ns with a motorized, computercontrolled linear stage. The delay interval between two consecutive data points was 13.3 fs for small delay times and was increased up to 200 ps for very large delay times. The stability of the samples was verified by recording the absorption spectra before and after the time-resolved measurements. The transient spectra recorded with the above outlined setup were free of interfering absorption signals in the ps time domain as has been detected for a number of aromatic solvents by Masuhara and Mataga.19 Furthermore, we did not observe any exciplex formation of NDI with toluene as was found by Barros et al.20 The original time-resolved spectra are given in the Supporting Information. These spectra are corrected for stray light and for the group velocity dispersion of the white light (chirp). The latter was done by fitting a polynomial to the cross phase modulation signal of the pure solvent under otherwise identical experimental conditions. The time-resolved spectra were analyzed by global fitting with GLOTARAN21 employing 15267

dx.doi.org/10.1021/jp304391x | J. Phys. Chem. C 2012, 116, 15265−15280

The Journal of Physical Chemistry C

Article

other regioisomers of the product mixture was achieved by recrystallization. The NDI-acceptor fragment 11 (see Scheme 2) was synthesized by an one-pot double imide formation in a

The stability of the samples was verified by recording the absorption spectra before and after the time-resolved measurements. As in the fs-experiments no transient absorption of toluene could be observed as has been detected for a number of aromatic solvents by Masuhara and Mataga.19 For the lifetime fits a parallel model (independent exponential decays) was assumed. Long lifetimes (>100 ns) were determined by a tail fitting routine of the decay curves while shorter ones (