Unveiling Singlet Fission Mediating States in TIPS-pentacene and its

May 22, 2015 - Unveiling Singlet Fission Mediating States in TIPS-pentacene and its Aza Derivatives ... E-mail: [email protected]. ...
0 downloads 12 Views 1MB Size
Download PDF
Article pubs.acs.org/JPCA

Unveiling Singlet Fission Mediating States in TIPS-pentacene and its Aza Derivatives Julia Herz,† Tiago Buckup,† Fabian Paulus,‡ Jens U. Engelhart,‡ Uwe H. F. Bunz,‡,§ and Marcus Motzkus*,†,§ †

Physikalisch-Chemisches Institut, Im Neuenheimer Feld 229, ‡Organisch-Chemisches Institut, Im Neuenheimer Feld 270, and Centre of Advanced Materials, Im Neuenheimer Feld 225, Ruprecht-Karls-Universität, Heidelberg D-69120, Germany

§

S Supporting Information *

ABSTRACT: Femtosecond pump−depletion−probe experiments were carried out in order to shed light on the ultrafast excited-state dynamics of triisopropylsilylethynyl (TIPS)-pentacene and two nitrogen-containing derivatives, namely, diazaTIPS-pentacene and tetraaza-TIPS-pentacene. Measurements performed in the visible and near-infrared spectral range in combination with rate model simulations reveal that singlet fission proceeds via the extremely short-lived intermediate 1TT state, which absorbs in the near-infrared spectral region only. The T1 → T3 transition probed in the visible region shows a rise time that comprises two components according to a consecutive reaction (S1 → 1TT → T1). The incorporation of nitrogen atoms into the acene structure leads to shorter dynamics, but the overall triplet formation follows the same kinetic model. This is of particular importance, since experiments on tetraaza-TIPS-pentacene allow for investigation of the triplet state in the visible range without an overlapping singlet contribution. In addition, the pump−depletion−probe experiments show that the triplet absorption in the visible (T1 → T3) and near-infrared (T1 → T2) regions occurs from the same initial state, which was questioned in previous studies. Furthermore, an additional ultrafast transfer between the excited triplet states (T3 → T2) is identified, which is also in agreement with the rate model simulation. By applying depletion pulses, which are resonant with higher vibrational levels, we gain insight into internal vibrational energy redistribution processes within the triplet manifold. This additional information is of great relevance regarding the study of loss channels within these materials.

■

INTRODUCTION Singlet fission (SF) has been attracting strong scientific attention due to its potential to raise the quantum efficiency of single junction solar cells. In this overall spin-conserved process, two molecules in the triplet state T1 are formed by investing just one photon.1−7 Pentacene undergoes SF efficiently and on an ultrafast time scale.8−10 External quantum efficiencies above 100% can be achieved in photovoltaic cells that use pentacene as a donor molecule.11 Triisopropylsilylethynyl (TIPS)-pentacene is an important, soluble derivative of pentacene with favorable intermolecular orientation within the π-stacked array.12,13 The excited-state dynamics of TIPSpentacene are widely studied by ultrafast transient absorption measurements.14−18 They have been described by a decay of the singlet state with a concomitant rise of the triplet species confirming SF as the underlying process. These studies concordantly show that the triplet excited-state absorption (ESA) is highly superimposed by the singlet ESA in the visible spectral region. The shoulder at higher energies of this signal (λ < 500 nm) is assigned to the (S1 → Sn) transition and the main peak around 530 nm to the triplet ESA (T1 → Tn). On the other hand, under the assumption that the triplet ESA in the © 2015 American Chemical Society

near-infrared (NIR) region is spectrally well separated, the triplet formation can be investigated independently from a singlet contribution. Accordingly, measurements on polycrystalline, unsubstituted pentacene8,9 and TIPS-pentacene films16 were carried out in this spectral region; however, in a recent study on transient absorption data of TIPS-pentacene, a global target analysis revealed also a small singlet contribution in the NIR region.17 Interestingly, the triplet ESA in the NIR region rises with a slightly shorter time constant compared to the visible region. According to the global target analysis, a kinetic model that describes the dynamics based on two separated T1 states in the two different spectral regions may explain the experimental observations. In order to gain more insight into the triplet population and relaxation dynamics, femtosecond pump−depletion−probe experiments in the visible region as well as in the NIR spectral range were performed. The idea of this technique is to set a specific marker within the kinetic network on a preselected Received: March 6, 2015 Revised: May 21, 2015 Published: May 22, 2015 6602

DOI: 10.1021/acs.jpca.5b02212 J. Phys. Chem. A 2015, 119, 6602−6610

Article

The Journal of Physical Chemistry A

of the depletion on the pump pulse is recorded. Within this study, we only show transient signals where the pump is chopped and delayed (by traveling through a delay stage). All three beams were focused onto the substrate with the sample under normal air conditions. Figure 1a presents the excitation

state that intercepts the normal relaxation dynamics. Experimentally, this is realized by introducing a second, so-called depletion or dump pulse that modifies the population after the initial excitation and thus controls the evolution of excited-state species.19−22 The changed population pathway can in turn identify specific states of the energy flow network that are dependent on the additional pulse. By combining information obtained from the experiment with simulations, we provide better understanding on how the triplet manifold is populated via the SF mechanism. In the present work, we want to reveal the origin of the different triplet rise times in the visible and NIR spectral regions observed in TIPS-pentacene and its aza derivatives.17 These new materials, which show similar SF dynamics, but spectrally shifted transitions, provide complementary information on the excited-state dynamics. This is shown for two compounds diaza- and tetraaza-TIPS-pentacene where two and four carbon atoms, respectively, are substituted by nitrogen atoms. As a consequence of the changed electronic properties in tetraaza-TIPS-pentacene, the singlet contribution is spectrally separated from the triplet ESA. Therefore, we can study the triplet formation almost independently from a singlet contribution at the wavelengths chosen for the pump− depletion−probe experiment.

Figure 1. (a) Energy diagram of TIPS-pentacene. The arrows illustrate the effect of the pump and the depletion pulses. (b) Top: pulse sequence in time; bottom: geometry of the three beams during the experiment.

■

scheme within the energy levels of TIPS-pentacene. The signal beam is focused on a silicon (visible) or InGaAs (NIR) photodiode array multichannel detector (256 pixels and a resolution of 1.32 nm per pixel). The pulse sequence is also shown together with the beam geometry of the pumpdepletion-probe experiment in Figure 1b. The pump−probe delay is defined as τprobe. After the pump pulse has promoted population into the excited state, the depletion pulse is applied with a variable delay T. Because of the fact that the depletion pulse is also partly resonant with the S0 → S1 transition in the case of TIPSpentacene, we could find temporal overlap of the pulses using the depletion as the excitation pulse. The time zero between the pump and depletion pulse was set indirectly after defining the temporal overlap first between the pump and the probe and second between the depletion and the probe. The spatial overlap between all three beams was optimized by maximizing the depletion effect at fixed T = 1 ps and fixed τprobe = 2 ps. The pump beam centered at 600 nm showed a pulse duration of 14 fs. The depletion pulse around 530 nm with a pulse duration of 18 fs has a bandwidth that nicely matches the triplet ESA of TIPS-pentacene as shown in Figure 2a. For tetraaza-TIPS-pentacene (Figure 2b), the bands are spectrally shifted resulting in a depletion spectrum that is not overlapped with a singlet contribution as discussed later. Both beams were attenuated to pulse energies of 40 nJ with a focal spot diameter of approximately 200 μm (ca. 4 × 1013 photons/ cm2) in order to minimize exciton−charge and exciton−exciton annihilation effects. The temporal resolution of the experiments was calculated (fitting curves of the coherent artifacts are shown in Figure S1, Supporting Information) to be 40 fs in the visible region and ∼90 fs in the NIR region.

EXPERIMENTAL SECTION All films were prepared on polyimide-coated 1737F glass substrates (PGO, Iserlohn, Germany). The ∼30 nm thick polyimide interlayer (diluted PI-2525, Hitachi Chemical DuPont MicroSystems GmbH, Neu-Isenburg, Germany) was used to achieve a good wettability of the organic solution and to provide a homogeneous film formation and crystallization. After cross-linking the polyimide film at 300 °C for 3 h, all substrates were rinsed in acetone and isopropyl alcohol in an ultrasonic bath for 5 min and dried with compressed air. TIPSpentacene (Sigma−Aldrich, Germany) was used without any further purification. Diaza-TIPS-pentacene and tetraaza-TIPSpentacene were synthesized as described in Engelhart et al.23 and Miao et al.,24 respectively. Both compounds were purified through recrystallization from hexane and ethanol. The small molecules were dissolved in anhydrous toluene (20 mg/mL, 10 mg/mL for tetraaza-TIPS-pentacene) and filtered through 0.45 PTFE filters prior to spin-coating. Films were spin-cast at 1500 rpm, followed by annealing (drying) at 40 °C for 30 min under nitrogen atmosphere (O2 and H2O < 1 ppm). Polarized microscope images were taken on a Nikon LV100-50i-pol under crossed polarizers indicating polycrystalline films. Steadystate optical VIS absorption spectra were recorded with a commercial UV−visible (UV−vis) spectrometer (Shimadzu 1800). For the ultrafast measurements, we used the output of a regeneratively amplified Ti/Sa laser generating 100 fs pulses with a repetition rate of 1 kHz. The fundamental laser was divided into two parts: 1% of the intensity was used to generate white light as the probe beam, and 99% was again split into two, resulting in a pump and a depletion beam. The white light was either made in a 2 mm sapphire crystal (450−750 nm) or in a 3 mm YAG crystal (830−1030 nm). Pump and depletion pulses were created in noncollinear optical parametric amplifiers in order to set the required wavelengths, respectively. Both the pump and the depletion pulse were compressed using prism compressors. Every second pump beam is blocked with a synchronized chopper wheel in order to get difference spectra, that is, with and without excitation. In this way, only the effect

■

RESULTS AND DISCUSSION We first present pump−depletion−probe measurements on TIPS-pentacene. A detailed rate model simulation that supports our experimental findings is shown subsequently. The outcome of this study motivates further scrutinizing of the fundamental steps of the SF process. A consecutive reaction was simulated to describe the different rise times in the visible and NIR spectral 6603

DOI: 10.1021/acs.jpca.5b02212 J. Phys. Chem. A 2015, 119, 6602−6610

Article

The Journal of Physical Chemistry A

Figure 2. UV−vis spectra: (a) spin-coated TIPS-pentacene thin film (black line) together with the excitation spectrum (red), the depletion spectrum (green), and the transient absorption spectrum at a pump−probe delay of 1 ps (gray shaded area); (b) tetraaza-TIPS-pentacene thin film.

regions. Furthermore, we focus on tetraaza-TIPS-pentecene, which represents a perfect system on which to apply pump− depletion−probe experiments, due to its spectrally separated singlet and triplet ESA bands. A thorough discussion about the derived kinetic model is given in the end. Pump−Depletion−Probe Experiments on TIPS-pentacene. The experimental results of the pump−depletion experiments probed in the visible region as well as in the NIR spectral region are shown in Figure 3. The excitation wavelength was at 600 nm, and the depletion pulse was located at 530 nm. As the pump beam is chopped, we just see an effect on the dynamics induced by the pump pulse, and the dynamics generated by the depletion pulse (S0 → S1) are not monitored. We first take a look on the rising dynamics of the triplet species in the visible spectral region. The transients illustrate the evolution of the ESA signals at 530 nm, which are disturbed by the depletion pulse at a selected time T. A loss of signal can be clearly seen when the depletion pulse comes at least 100 fs after the pump pulse (when T = 100 fs, 200 fs, and 1 ps). The resulting dip feature is sharp and more pronounced at larger T values. The signal recovers after the perturbation, but not all of the population comes back, when compared to the signal intensity before the depletion pulse arrives. In spite of an experimental response time of approximately 40 fs, the typical dip feature caused by the interaction with the depletion pulse cannot be resolved for T = 50 fs (Figure 3a). Here, the depletion effect is overlapped with the coherence spike, resulting from the interaction between the pump and the probe pulse. There is a depletion effect from the beginning on when compared to a transient where no depletion pulse is applied: however, it is challenging to quantify the interaction (for further explanation see the Supporting Information and Figure S2). The experimental data probed in the NIR region look different. A loss of signal after depleting the triplet ESA in the visible region is clearly visible only for T = 1 ps. However, the dip feature is less sharp and builds up with time. No recovery of the signal is measured after the depletion action, but a decay behavior can be observed instead, or in another respect, a small rise of the depletion effect. At T = 200 fs, a depletion effect can still be detected, but the dip feature is much less pronounced, and the signal does not recover again. For T values of 100 and 50 fs, no dip feature is observed, within our given temporal resolution of ∼90 fs. The experimental results in the visible spectral range are explained as follows. Considering a time constant of 150 fs for

Figure 3. Results of the pump−depletion−probe experiments probed in (a) the visible spectral region and (b) the NIR spectral region. The probing wavelength of the depicted transients is 530 nm in (a) and an averaged wavelength range from 850 to 860 nm in (b). The data sets were vertically shifted in order to clearly illustrate the depletion effect of the single transients. The arrows indicate the arrival time of the depletion pulse.

6604

DOI: 10.1021/acs.jpca.5b02212 J. Phys. Chem. A 2015, 119, 6602−6610

Article

The Journal of Physical Chemistry A the triplet rise (obtained from a global fit, see Figure S6, Supporting Information, and ref 17), at τprobe = 1 ps, all of the population is already transferred to the triplet state.17 As we apply the depletion pulse on the triplet ESA, the depletion effect increases with time and is most pronounced at T = 1 ps (the maximum T value within this study) (Figure 3a). However, we do see a depletion effect in the transient data at early time delays (τprobe = 50−200 fs) when the triplet state has not fully developed. Thus, the observation of a dip feature when the depletion pulse interacts 100 and 200 fs after the pump pulse indicates that also the singlet state (S1 → Sn transition) is affected by the depletion pulse centered at 530 nm (see Figure S6, Supporting Information). This finding is related to the broad spectral overlap of the singlet and triplet ESAs in the visible spectral region. Since the depletion pulse is also resonant with the ground-state spectrum of TIPS-pentacene, it re-excites the S1 state (Figure S5, Supporting Information), which is what decreases the effect of the triplet ESA depletion. The signal growth (or decay of the depletion effect) after the depletion action in the visible can be explained by this re-excitation, which later repopulates the triplet manifold. From the transient probed at 530 nm, we can estimate, by looking at the signal before and after the depletion action, that approximately onehalf of the population comes back. The time constants of the recovery are for all measured transients (T = 50 fs, 100 fs, 200 fs, and 1 ps) in the same range (200−250 fs, see the fitting curves in Figure S3 and Table S1, Supporting Information) when compared to the undisturbed triplet rise. The experimental data show that the triplet ESA probed in the NIR region is influenced by the depletion pulse applied in the visible region (Figure 4). The different evolution of the signal

In relation to the excited-state dynamics of TIPS-pentacene, more population in the excited state (T2) or less population in the T1 state is required to get the inverted recovery signal. This would either require a decay of the triplet state T1, which is highly unlikely (and not observed in the visible data), or a population transfer to the T2 state. A rate model simulation is performed to clarify if the slow increase of the depletion effect (NIR) can be generated by a population gain in the excited state that is probed (T2) in contrast to the sharp dip feature seen in the visible region. The interpretation of the data is only possible if all the mentioned contributions to the signal are considered and quantified. In other words, the simulation of the depletion signal, to extract information from the experimental results, becomes essential. Simulation of the Depletion Signal. The input for the rate model simulation consists of the population of the singlet ground (S0) and excited states (S1), the population of the triplet states (Thot, T1, T2, T3), and their connected decay rates (see the rate model in Figure S7 and Table S3, Supporting Information). In principle, the population of each state is calculated dependent on the decay rate for every delay time τprobe. At a set time T, the depletion effect is applied. The amount of population that gets depleted can be obtained from the amount of the depletion effect (dip) in the experimental data. Since the depletion at 530 nm is resonant with the T1 → T3 transition (ref 19), the excitation of T3 is also included. Furthermore, a certain amount of singlet repopulation due to the S0 → S1 excitation by the depletion pulse itself is considered. To simulate the signal at a specific wavelength, the relative amplitudes of the singlet and triplet contributions have to be considered. For example, at 530 nm, the amount of the triplet ESA contribution relative to the singlet ESA contribution is about 80%. The presented figures take the populations of S1, Thot, T1, T2, T3, and the population difference T1 − T3 (T1 − T2) into account. In order to consider the coherence spike, the nonlinear optical interaction was also simulated according to refs 25 and 26 (see eq 1, Supporting Information). Additionally, we convolute our simulated signal with a Gaussian to take our experimental resolution into account. Figure 5a shows the simulation of the visible pump− depletion−probe data. The time resolution was set to be 40 fs. The rise of the signal shows the population of the triplet state. At T = 1 ps, the depletion pulse is applied. The resulting sharp dip feature, seen in the experimental data (Figure 3a), is well reproduced in the simulation. The signal recovery afterward can be explained by the triplet repopulation resulting from the singlet re-excitation as mentioned before. In correspondence with our experimental settings, the temporal resolution of the simulated NIR data is lower, which becomes apparent in the broad coherence spike (Figures 3b, 5b, and S1b, Supporting Information). This leads to a temporal broadening of the dip feature in accordance with the experimental observations. As mentioned before, the signal decay after the depletion action instead of a recovery (as found for the visible data) can only occur if the final state of a transition gains population via an additional transfer to the T2 state. If a population transfer from the T3 to the T2 state is included in the rate model simulations, the experimental observations can perfectly be reproduced. This implies, that an additional ultrafast T3 → T2 transition (ca. 90 fs) is required, which seems feasible, since the T3 state lies energetically higher than the T2 state.27 The fact that the NIR ESA is affected by the depletion pulse indicates that both transitions (visible and NIR) have

Figure 4. Comparison of the depletion effect (depletion at 530 nm, 1 ps after excitation) probed in the visible (530 nm, black curve) and the NIR (850−860 nm, red curve) spectral regions for TIPS-pentacene. Data was normalized at τprobe = 2 ps.

after applying the depletion pulse at T = 1 ps is particularly puzzling. If the transition in the NIR and visible regions would occur from the same state, the depletion pulse should have the same effect on the dynamics, that is, lead to a sharp dip feature and a signal recovery after the depletion action. In order to observe a signal evolution after the depletion action as experimentally found for the NIR ESA (Figure 4), that is, a further decay of the signal, a population gain in the excited or a population loss in the initial state is required. 6605

DOI: 10.1021/acs.jpca.5b02212 J. Phys. Chem. A 2015, 119, 6602−6610

Article

The Journal of Physical Chemistry A

formation of the triplet band in the visible (150 fs) and the NIR (120 fs) spectral regions, respectively, give access to decipher the entangled dynamics at the initial stages of SF. There is already some general agreement that formation of the free triplets occurs via an intermediate coupled triplet state (1TT).1,2,16,28−30 This correlated triplet exciton pair is termed a multiexciton (ME)-state intermediate10,30,31 or an optically dark (D) multiexciton,32 which couples nonadiabatically to the bright S1 state. As its lifetime is believed to be extremely short, it has been very challenging until now to detect this species by means of conventional transient absorption measurements due to temporal and spectral overlap of several contributions at early time delays, for example, coherent effects, and singlet and triplet absorption bands. By taking 1TT into account and implementing it into the rate model via a consecutive reaction (S1 → 1TT → T1), the different experimental triplet rise times in the NIR and visible regions can now be perfectly justified (Figure 6); while the S1 and T1 states absorb in the visible

Figure 6. Evolution of species in a consecutive reaction with t1 = 50 fs and t2 = 135 fs. If the detection pulse probes both states (NIR), a sum of the 1TT and the T1 states (with equal weighting) can be seen, resulting in a rise time of 120 fs. If the detection is specific only to T1 (visible), the rise time comes to 155 fs. The fitting is done with a single exponential growth for both curves.

Figure 5. Simulation of the pump−depletion−probe signal of TIPSpentacene. Dashed black line shows the transient dynamics without depletion effect. Red curve demonstrates the dynamics after applying a depletion pulse at 1 ps. Dynamics in the (a) visible and (b) NIR spectral ranges. The energy diagrams present the underlying processes that cause the signal. (c) Zoom into the signal in the NIR region, and the blue curve takes the experimental response time into account. Inset shows the simulated signal when the depletion pulse arrives 100 fs after the pump pulse with an experimental time resolution of 35 fs.

region as well as in the NIR region (see SAS spectra in Figure S6, Supporting Information), the 1TT state just absorbs in the NIR region. This leads to overlapping contributions, which results in slightly different experimental rise times in the visible and NIR regions. According to this observation, the simulation leads to a time constant of 50 fs for the rise of the coupled triplet pair state 1 TT, and the subsequent step, the formation of the T1 state, occurs in 135 fs. In the visible region, a superposition of the two time constants adds up to 155 ± 10 fs. In the NIR spectral range, however, the situation is different. In this spectral region, also, the 1TT species absorbs, and the measured signal exhibits a shorter rise time constant of 120 fs. Furthermore, a closer look into the evolution of the signal in the NIR region shows that the rising times increase with higher wavelengths (from 850 to 900 nm, see Figure S4, Supporting Information, and rise times in Table S2, Supporting Information). This trend originates from contributions with different signal amplitudes, which are wavelength dependent. The 1TT state has a larger amplitude at shorter wavelengths and thus exceeds the absorption signal of the T1 state. At wavelengths between

their origin in the same initial state T1. However, at first glance, one would expect that the same sharp dip feature should be observed also in the NIR region. A zoom of the simulation of this depletion action is presented in Figure 5c that explains the different shape. The experimental response time (blue curve, Figure 5c) is much longer so that the depletion effect cannot be resolved. The limited temporal resolution in the NIR region also explains the absence of a dip feature at very early delay times. In order to observe a small dip at a depletion time of T = 100 fs, a temporal resolution of about 35 fs is required as shown in the simulation (inset Figure 5c). Assignment of Rise Times in the Visible and NIR Spectral Regions. The experimental results and the outcome of the rate model simulation clearly show that one single triplet state T1 is populated from the singlet manifold via SF. The observation of two different rise times observed for the 6606

DOI: 10.1021/acs.jpca.5b02212 J. Phys. Chem. A 2015, 119, 6602−6610

Article

The Journal of Physical Chemistry A 850 and 950 nm, the 1TT contribution is smaller and superimposed by T1. Therefore, it is not possible to see 1TT directly in transient absorption using probe pulses above 840 nm and experimental response times greater than 20 fs. Because of the spectral proximity of the S1 and 1TT states, the vibronic coupling between them should be observable using multidimensional techniques such as two-dimensional electronic spectroscopy30,33 or pump-degenerate four-wave mixing.34 It is important to note that this intermediate state does not influence the dynamics of the depletion effect at 1 ps. Hence, the simulation of the depletion effect shown before is not affected by this new finding. Pump−Depletion−Probe Experiments on TetraazaTIPS-pentacene. In all of the pump−depletion−probe experiments on TIPS-pentacene that we have shown so far, the depletion pulse not only affected the triplet manifold, but the S1 state was also influenced. An independent depletion action on the T1 state is not feasible in TIPS-pentacene, since a depletion pulse around 530 nm is always resonant also with the S1 → Sn transition. To separate this excitation and investigate the effect of a depletion pulse on the triplet state only, we focus on an aza derivative of TIPS-pentacene (tetraaza-TIPSpentacene) that does not exhibit a singlet contribution in the region where the depletion pulse is applied (530 nm). The transient absorption data of tetraaza-TIPS-pentacene, which exhibits two pyrazine units symmetrically incorporated into the TIPS-pentacene backbone, are shown in Figures S11−S13 (Supporting Information). The steady-state absorption spectrum together with pump and depletion pulses and chemical structure are depicted in Figure 2b. SF dynamics are very similar to those of TIPS-pentacene; however, the relevant states are different in energy. The HOMO−LUMO transition is redshifted, due to the lower frontier orbitals compared to TIPSpentacene.24 The T1 → T3 transition is also shifted to higher wavelengths and peaks around 560 nm. From the global analysis, we obtain a first species around 610 nm, which is attributed to a state with singlet character. This species decays with a time constant of 100 fs toward the triplet manifold (see Figure S13, Supporting Information). Altogether, tetraazaTIPS-pentacene behaves similar to TIPS-pentacene; however, the singlet species has negligible amplitude at 530 nm, where the depletion spectrum is located. Thus, the depletion pulse does not affect the singlet state (S1−Sn). This can be nicely seen in Figure 7. At early time delays when the system is still in the excited singlet state, the depletion has no effect on the dynamics as can be seen for short T values (T = 50 and 100 fs). The dip feature is first observable at T = 200 fs, when the triplet state has already developed. A strong depletion effect on the triplet state is visible at T = 1 ps. However, compared to the same measurements on TIPS-pentacene, the dip feature is less sharp, and only a small amount of population returns to the T1 state. The signal evolution looks similar to the depletion effect observed for TIPS-pentacene in the NIR region (compare Figure 7 with Figure 3). In comparison to TIPS-pentacene, the triplet T1 → T3 transition is red-shifted (around 560 nm), and higher lying vibrational states are addressed by the depletion pulse at 530 nm. The excitation of the population, blue-shifted with regard to the lowest transition, allows for a detailed discussion of the depletion effect. In Figure 8, selected wavelengths are depicted in order to compare the effect of the depletion pulse at 530 nm on the dynamics at different positions in the spectrum.

Figure 7. Results of the pump−depletion−probe experiments of tetraaza-TIPS-pentacene, pumped at 600 nm, depleted at 530 nm, and probed in the visible spectral region (an averaged wavelength range from 555 to 565 nm). The arrows indicate the arrival time of the depletion pulse.

Figure 8. Pump−depletion−probe data of tetraaza-TIPS-pentacene after applying the depletion pulse at 530 nm, 1 ps after the pump pulse at selected probing wavelengths (530−570 nm).

If we probe at wavelengths where the depletion pulse acts, we can directly monitor intramolecular vibrational energy redistribution (IVR) back to the lowest vibrational state in T3. Population loss in the T3hot state (vibrationally excited T3 state) reduces the depletion effect when exactly this state is probed. In another respect, the population gain of the lowest vibrational T3 state results in a further decay of the signal on the time scale of IVR. Signal recovery, resulting from a repopulation of the T3 state as a consequence of re-exciting S1 is also not observable when probing higher vibrational levels in T3. At higher wavelengths, in this case the maximum of the T1 → T3 absorption (560 nm), we do see a clear dip feature, which is slightly delayed. This is related to the fact that we probe lower vibrational levels in T3 than were originally excited by the depletion pulse. The depletion effect increases while the state, which is probed, gains population. As confirmed by our rate model simulation, we can resemble a further decay after the depletion action if a T3hot → T3 transition is included (for simulation, kinetic model, and rates see Figures S18 and S19 and Table S4, Supporting Information). We do also observe a signal recovery; however, this repopulation of T3 does not possess the same time constant as the initial triplet rise 6607

DOI: 10.1021/acs.jpca.5b02212 J. Phys. Chem. A 2015, 119, 6602−6610

Article

The Journal of Physical Chemistry A

Figure 9. Detailed kinetic model for the excited-state dynamics of TIPS-pentacene derived from pump−depletion−probe experiments and simulations on the left. For comparison, the respective time constants of the aza derivatives are listed in the table on the right.

exciton state with a time constant of 135 ± 10 fs in the case of TIPS-pentacene (for comparison, the time constants of the aza derivatives are shown within a table in Figure 9). In the visible spectral region, only the T1 → T3 ESA is observed, which rises with a time constant resulting from a sum of 50 and 135 fs. The lifetime of T1 exceeds the observation time of the experiment (1 ns). Furthermore, an additional ultrafast transition of ∼90 fs from the T3 to the T2 state is necessary to explain the pump− depletion signal probed in the NIR spectral region. When probing at wavelengths close to the energy of the depletion pulse, which is resonant with higher lying vibrational states, IVR processes in the T3 state can be monitored directly. Recently, Kolata et al. stated that, after 1TT formation, an excimer-like (EX) state is formed as an intermediate rather than a competitive state in perfluoropentacene.35 In this model system, a 12 ps time constant is assigned to the relaxation of EX and the dephasing of 1TT. According to their energy level scheme, EX lies close to 1TT. Since the species-associated spectra of the second species (gained from our global analysis on TIPS-pentacene and aza derivatives, see Figure S14, Supporting Information, and ref 17) looks almost identical to the final relaxed triplet state, an assignment of this species to the EX state is not possible. However, the formation of two individual triplet states includes decoherence and relaxation of the 1TT state as well as diffusive processes. Since these mechanisms are on larger time scales than 100−150 fs, the requirement of an intermediate state seems to be likely and is considered as an additional relaxation time constant of ∼1 ps within the T1 state. Most probably, this time constant is connected to an IVR within the vibrational manifold of the T1 state.

(compare signal evolution before and after the depletion action in Figure 8). Both, the repopulation and IVR transfers population into the lowest vibrational T3 state. Therefore, the signal recovery is reduced, since it is compensated by IVR. At intermediate wavelengths (540−550 nm), we observe a behavior that lies in between, where the depletion effect is superimposed by IVR. This leads to a signal evolution, which shows a combination of both depending on the amount of each contribution as presented in Figure 8 for the transients at 550 and 560 nm. Transients of the pump−depletion experiment of tetraaza-TIPS-pentacene probed in the NIR region for T = 200 fs, 500 fs, and 1 ps are shown in Figure S14 (Supporting Information). The effect of the elongated decay after the depletion action is even more dramatic in tetraaza-TIPSpentacene compared to TIPS-pentancene in the NIR region. This is intuitive, since T3hot first decays into the relaxed T3 state via IVR before the T3 → T2 transition can take place. We also applied the pump−depletion−probe experiment on diaza-TIPS-pentacene, which exhibits one pyrazine unit in the TIPS-pentacene backbone. It has already been shown that SF is accelerated in this material most likely due to enhanced intermolecular coupling.17 Faster SF as a consequence of reduced intermolecular distance was also observed recently for a similar nitrogen-containing compound.18 The dynamics after the depletion pulse are similar to those of TIPS-pentacene, because diaza-TIPS-pentacene also exhibits a singlet transition that is effected by the depletion pulse at 530 nm. This is illustrated in Figure S15 (Supporting Information) where a dip feature is observable at T = 100 fs. This dip feature seems to be delayed, since the depletion pulse is resonant with higher vibrational levels. The result of this T3hot population is discussed in detail for tetraza-TIPS-pentacene (see above). Because of our restricted experimental response, we could not detect a depletion effect at T = 100 fs in the NIR spectral region (see Figure S16, Supporting Information). Combining the information on the investigated systems, we can provide a detailed picture on the relaxation dynamics after photoexcitation of TIPS-pentacene and the two aza derivatives. In Figure 9, we present the conclusion of our study within a kinetic model of the excited-state dynamics. Excitation into the first excited state is followed by an ultrafast transition to the coupled triplet pair state 1TT. This state absorbs in the NIR region but cannot be detected directly due its extremely short lifetime and broad spectral distribution. It decays to the triplet

■

CONCLUSION Experiments on the excited-state dynamics on TIPS-pentacene and the two aza derivatives, diaza-TIPS-pentacene and tetrazaTIPS-pentacene, reveal that the formation of a single triplet state occurs via the intermediate coupled triplet pair state 1TT for all three compounds. The nitrogen substitution in the aza derivatives does not just accelerate the formation of the 1TT state but also the formation of the triplet state T1 from the 1TT state. 1TT builds up within 25 fs for diaza-TIPS-pentacene and within 35 fs for tetraaza-TIPS-pentacene, which is faster than the time constant obtained for TIPS-pentacene (50 fs). Moreover, our experimental findings show that 1TT absorbs 6608

DOI: 10.1021/acs.jpca.5b02212 J. Phys. Chem. A 2015, 119, 6602−6610

Article

The Journal of Physical Chemistry A

et al. Multiphonon Relaxation Slows Singlet Fission in Crystalline Hexacene. J. Am. Chem. Soc. 2014, 136, 10654−10660. (8) Wilson, M. W. B.; Rao, A.; Clark, J.; Kumar, R. S. S.; Brida, D.; Cerullo, G.; Friend, R. H. Ultrafast Dynamics of Exciton Fission in Polycrystalline Pentacene. J. Am. Chem. Soc. 2011, 133, 11830−11833. (9) Rao, A.; Wilson, M. W. B.; Hodgkiss, J. M.; Albert-Seifried, S.; Bässler, H.; Friend, R. H. Exciton Fission and Charge Generation via Triplet Excitons in Pentacene/C60 Bilayers. J. Am. Chem. Soc. 2010, 132, 12698−12703. (10) Chan, W.-L.; Ligges, M.; Jailaubekov, A.; Kaake, L.; Miaja-Avila, L.; Zhu, X.-Y. Observing the Multiexciton State in Singlet Fission and Ensuing Ultrafast Multielectron Transfer. Science 2011, 334, 1541− 1545. (11) Congreve, D. N.; Lee, J.; Thompson, N. J.; Hontz, E.; Yost, S. R.; Reusswig, P. D.; Bahlke, M. E.; Reineke, S.; Van Voorhis, T.; Baldo, M. A. External Quantum Efficiency above 100% in a Singlet-ExcitonFission−Based Organic Photovoltaic Cell. Science 2013, 340, 334−337. (12) Anthony, J. E.; Brooks, J. S.; Eaton, D. L.; Parkin, S. R. Functionalized Pentacene: Improved Electronic Properties from Control of Solid-State Order. J. Am. Chem. Soc. 2001, 123, 9482−9483. (13) Anthony, J. E.; Eaton, D. L.; Parkin, S. R. A Road Map to Stable, Soluble, Easily Crystallized Pentacene Derivatives. Org. Lett. 2001, 4, 15−18. (14) Ramanan, C.; Smeigh, A. L.; Anthony, J. E.; Marks, T. J.; Wasielewski, M. R. Competition between Singlet Fission and Charge Separation in Solution-Processed Blend Films of 6,13-Bis(triisopropylsilylethynyl)pentacene with Sterically-Encumbered Perylene-3,4:9,10-bis(dicarboximide)s. J. Am. Chem. Soc. 2011, 134, 386− 397. (15) Walker, B. J.; Musser, A. J.; Beljonne, D.; Friend, R. H. Singlet Exciton Fission in Solution. Nat. Chem. 2013, 5, 1019−1024. (16) Yost, S. R.; Lee, J.; Wilson, M. W. B.; Wu, T.; McMahon, D. P.; Parkhurst, R. R.; Thompson, N. J.; Congreve, D. N.; Rao, A.; Johnson, K.; et al. A Transferable Model for Singlet-Fission Kinetics. Nat. Chem. 2014, 6, 492−497. (17) Herz, J.; Buckup, T.; Paulus, F.; Engelhart, J.; Bunz, U. H. F.; Motzkus, M. Acceleration of Singlet Fission in an Aza-derivative of TIPS-pentacene. J. Phys. Chem. Lett. 2014, 5, 2425−2430. (18) Wu, Y. S.; Liu, K.; Liu, H. Y.; Zhang, Y.; Zhang, H. L.; Yao, J. N.; Fu, H. B. Impact of Intermolecular Distance on Singlet Fission in a Series of TIPS Pentacene Compounds. J. Phys. Chem. Lett. 2014, 5, 3451−3455. (19) Wohlleben, W.; Buckup, T.; Hashimoto, H.; Cogdell, R. J.; Herek, J. L.; Motzkus, M. Pump-Deplete-Probe Spectroscopy and the Puzzle of Carotenoid Dark States. J. Phys. Chem. B 2004, 108, 3320− 3325. (20) Kee, T. W. Femtosecond Pump-Push-Probe and Pump-DumpProbe Spectroscopy of Conjugated Polymers: New Insight and Opportunities. J. Phys. Chem. Lett. 2014, 5, 3231−3240. (21) Bakulin, A. A.; Rao, A.; Pavelyev, V. G.; van Loosdrecht, P. H. M.; Pshenichnikov, M. S.; Niedzialek, D.; Cornil, J.; Beljonne, D.; Friend, R. H. The Role of Driving Energy and Delocalized States for Charge Separation in Organic Semiconductors. Science 2012, 335, 1340−1344. (22) Buckup, T.; Savolainen, J.; Wohlleben, W.; Herek, J. L.; Hashimoto, H.; Correia, R. R. B.; Motzkus, M. Pump-Probe and Pump-Deplete-Probe Spectroscopies on Carotenoids with N = 9−15 Conjugated Bonds. J. Chem. Phys. 2006, 125, 194505−194507. (23) Engelhart, J. U.; Lindner, B. D.; Tverskoy, O.; Rominger, F.; Bunz, U. H. F. Pd-Catalyzed Coupling of Non-activated Dibromoarenes to 2,3-Diaminoarenes: Formation of N,N′-Dihydropyrazines. Chem.Eur. J. 2013, 19, 15089−15092. (24) Miao, S.; Appleton, A. L.; Berger, N.; Barlow, S.; Marder, S. R.; Hardcastle, K. I.; Bunz, U. H. F. 6,13-Diethynyl-5,7,12,14-tetraazapentacene. Chem.Eur. J. 2009, 15, 4990−4993. (25) Lorenc, M.; Ziolek, M.; Naskrecki, R.; Karolczak, J.; Kubicki, J.; Maciejewski, A. Artifacts in Femtosecond Transient Absorption Spectroscopy. Appl. Phys. B: Lasers Opt. 2002, 74, 19−27.

in the NIR region only, while the singlet and triplet absorb in the visible and the NIR regions. The superposition of the 1TT and T1 ESA bands in the NIR region leads to an apparent faster triplet rise compared to the visible region, where only the T1 → T3 ESA is observed. Further knowledge about the interplay between excited states can help in a future step, to modify special transitions on a molecular level. In that way, namely, by avoiding or activating transitions, the SF process can be optimized. The potential to enhance and direct the SF mechanism by intercepting the normal relaxation dynamics is attractive, regarding the design of new materials and their application in organic photovoltaic devices.

■

ASSOCIATED CONTENT

* Supporting Information S

Experimental details, additional pump-depletion-probe data, fitting curves, rate models for simulation, transient absorption measurements on tetraaza-TIPS-pentacene, and sample characterization. The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/ acs.jpca.5b02212.

■

AUTHOR INFORMATION

Corresponding Author

* E-mail: [email protected]. Author Contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS A scholarship of the graduate college “Connecting Molecular πSystems into Advanced Functional Materials” provided by the BW-Landesgraduiertenförderung to J.H. is gratefully appreciated.

■ ■

ABBREVIATIONS USED SF, singlet fission; NIR, near-infrared; ESA, excited-state absorption; TIPS, triisopropylsilylethynyl REFERENCES

(1) Smith, M. B.; Michl, J. Singlet Fission. Chem. Rev. 2010, 110, 6891−6936. (2) Smith, M. B.; Michl, J. Recent Advances in Singlet Fission. Annu. Rev. Phys. Chem. 2013, 64, 361−386. (3) Roberts, S. T.; McAnally, R. E.; Mastron, J. N.; Webber, D. H.; Whited, M. T.; Brutchey, R. L.; Thompson, M. E.; Bradforth, S. E. Efficient Singlet Fission Discovered in a Disordered Acene Film. J. Am. Chem. Soc. 2012, 134, 6388−6400. (4) Burdett, J. J.; Bardeen, C. J. The Dynamics of Singlet Fission in Crystalline Tetracene and Covalent Analogs. Acc. Chem. Res. 2013, 46, 1312−1320. (5) Lee, J.; Bruzek, M. J.; Thompson, N. J.; Sfeir, M. Y.; Anthony, J. E.; Baldo, M. A. Singlet Exciton Fission in a Hexacene Derivative. Adv. Mater. 2013, 25, 1445−1448. (6) Ito, A.; Shimizu, A.; Kishida, N.; Kawanaka, Y.; Kosumi, D.; Hashimoto, H.; Teki, Y. Excited-State Dynamics of Pentacene Derivatives with Stable Radical Substituents. Angew. Chem., Int. Ed. 2014, 53, 6715−6719. (7) Busby, E.; Berkelbach, T. C.; Kumar, B.; Chernikov, A.; Zhong, Y.; Hlaing, H.; Zhu, X. Y.; Heinz, T. F.; Hybertsen, M. S.; Sfeir, M. Y.; 6609

DOI: 10.1021/acs.jpca.5b02212 J. Phys. Chem. A 2015, 119, 6602−6610

Article

The Journal of Physical Chemistry A (26) Dobryakov, A. L.; Kovalenko, S. A.; Ernsting, N. P. Coherent and Sequential Contributions to Femtosecond Transient Absorption Spectra of a Rhodamine Dye in Solution. J. Chem. Phys. 2005, 123, 044502−044508. (27) Pabst, M.; Kohn, A. Implementation of Transition Moments between Excited States in the Approximate Coupled-Cluster Singles and Doubles Model. J. Chem. Phys. 2008, 129, 214101−214112. (28) Kolomeisky, A. B.; Feng, X.; Krylov, A. I. A Simple Kinetic Model for Singlet Fission: A Role of Electronic and Entropic Contributions to Macroscopic Rates. J. Phys. Chem. C 2014, 118, 5188−5195. (29) Wang, L. J.; Olivier, Y.; Prezhdo, O. V.; Beljonne, D. Maximizing Singlet Fission by Intermolecular Packing. J. Phys. Chem. Lett. 2014, 5, 3345−3353. (30) Rao, A.; Bakulin, A.; Egorova, D.; Chin, A.; Zigmantas, D.; Morgan, S. Vibrational Coherence Reveals the Role of Dark Multiexciton States in Ultrafast Singlet Exciton Fission. OSA Tech. Dig. 2014, 11.Fri.A.4 DOI: 10.1364/UP.2014.11.Fri.A.4. (31) Chan, W.-L.; Berkelbach, T. C.; Provorse, M. R.; Monahan, N. R.; Tritsch, J. R.; Hybertsen, M. S.; Reichman, D. R.; Gao, J.; Zhu, X. Y. The Quantum Coherent Mechanism for Singlet Fission: Experiment and Theory. Acc. Chem. Res. 2013, 46, 1321−1329. (32) Zimmerman, P. M.; Musgrave, C. B.; Head-Gordon, M. A Correlated Electron View of Singlet Fission. Acc. Chem. Res. 2013, 46, 1339−1347. (33) Milota, F.; Prokhorenko, V. I.; Mancal, T.; von Berlepsch, H.; Bixner, O.; Kauffmann, H. F.; Hauer, J. Vibronic and Vibrational Coherences in Two-Dimensional Electronic Spectra of Supramolecular J-Aggregates. J. Phys. Chem. A 2013, 117, 6007−6014. (34) Buckup, T.; Motzkus, M. Multidimensional Time-Resolved Spectroscopy of Vibrational Coherence in Biopolyenes. Annu. Rev. Phys. Chem. 2014, 65, 39−57. (35) Kolata, K.; Breuer, T.; Witte, G.; Chatterjee, S. Molecular Packing Determines Singlet Exciton Fission in Organic Semiconductors. ACS Nano 2014, 8, 7377−7383.

6610

DOI: 10.1021/acs.jpca.5b02212 J. Phys. Chem. A 2015, 119, 6602−6610