Excited-State Dynamics of 5,14- vs 6,13-Bis(trialkylsilylethynyl)-Substituted Pentacenes: Implications for Singlet Fission.

Singlet fission is a process in conjugated organic materials that has the potential to considerably improve the performance of devices in many applications, including solar energy conversion. In any application involving singlet fission, efficient triplet harvesting is essential. At present, not much is known about molecular packing arrangements detrimental to singlet fission. In this work, we report a molecular packing arrangement in crystalline films of 5,14-bis(triisopropylsilylethynyl)-substituted pentacene, specifically a local (pairwise) packing arrangement, responsible for complete quenching of triplet pairs generated via singlet fission. We first demonstrate that the energetic condition necessary for singlet fission is satisfied in amorphous films of the 5,14-substituted pentacene derivative. However, while triplet pairs form highly efficiently in the amorphous films, only a modest yield of independent triplets is observed. In crystalline films, triplet pairs also form highly efficiently, although independent triplets are not observed because triplet pairs decay rapidly and are quenched completely. We assign the quenching to a rapid nonadiabatic transition directly to the ground state. Detrimental quenching is observed in crystalline films of two additional 5,14-bis(trialkylsilylethynyl)-substituted pentacenes with either ethyl or isobutyl substituents. Developing a better understanding of the losses identified in this work, and associated molecular packing, may benefit overcoming losses in solids of other singlet fission materials.

Introduction

Singlet fission is a process unique to molecular photophysics where one spin-singlet exciton is split into two spin-triplet excitons. The generally accepted kinetic scheme for singlet fission[1−3] is shown below:The correlated triplet pair, 1(TT) [abbreviated henceforth as TT for simplicity], is a critical intermediate in singlet fission. The correlated triplet pair is initially of overall singlet spin, meaning the process can occur on exceptionally rapid picosecond and subpicosecond time scales. Because competing decay processes are generally orders of magnitude slower, singlet fission can occur with exceptionally high efficiency, essentially splitting every photoexcitation into two independent triplet excitons lacking spin correlation (i.e., T1 + T1). Incorporating the singlet fission process into an optoelectronic device, such as a single-junction solar cell, could boost power-conversion efficiencies by as much as ∼1.4.[4] Such efficiency gains, however, require harvesting every triplet that is generated.

In moving toward practical applications, a thin singlet fission sensitizer layer is necessary. Within this sensitizer layer, intermolecular packing is of utmost importance. In crystalline material, for example, quantitative triplet yields (i.e., ΦT1 ∼ 200%) have been reported whereas in other material forms, such as in the amorphous phase, losses are generally observed (i.e., ΦT1 ≪ 200%).[5−8] In addition, in crystalline material both local molecular packing and long-range order can be determined via structural characterization, and most studies so far have emphasized crystalline material with packing arrangements optimal for singlet fission.[9−24] Solution-processable molecular semiconductors, such as functionalized acene derivatives, present themselves as a promising material system for this sensitization layer as they can often be deposited as uniform films and subsequently crystallized through various annealing techniques into the desired molecular packing.[25−28]

To date, there exists only one general design rule for molecular packing—specifically, that some displacement is necessary along the direction of the intermolecular axis associated with the lowest singlet excited-state transition to maximize the rate of the first step in singlet fission.[1,2] For most acene derivatives, this optimal displacement is along the short molecular axis. Empirically, we have shown that crystalline solids (i.e., films and nanoparticles) of a number of pentacene derivatives exhibit highly efficient singlet fission.[6,29,30] These compounds tend to adopt a packing arrangement with some displacement along the short axis and more extensive displacement along the long axis (for more on this, see below). That said, far fewer studies, if any, have reported molecular packing that can be detrimental to singlet fission.

In this article, we report a local (pairwise) packing arrangement in crystalline films of 5,14-bis(triisopropylsilylethynyl)pentacene (5,14-TIPS-Pn) that results in complete quenching of triplet pairs generated via singlet fission. Singlet fission takes place when the compound is in its amorphous phase, absent of a common packing arrangement. While triplet pairs form in high yield in crystalline films, we show that strong coupling to the ground state promotes rapid quenching of the triplet pairs, resulting in quantitative triplet losses (ΦT1 ∼ 0%). Similar quantitative quenching of triplet pairs is observed in crystalline films of two additional 5,14-substituted pentacene derivatives with ethyl and isobutyl substituents. Developing a better understanding of the losses identified in this work and associated molecular packing, which may represent an important loss pathway in solids of pentacene derivatives, motivates deeper examinations along with studies of similar molecular pair arrangements in other singlet fission materials.

Experimental Methods

Pentacene Derivative Synthesis

The synthesis of 5,14-bis(trialkylsilylethynyl)pentacenes has been reported previously.[31,32]

Single Crystal Structures

The crystal structures of 5,14- and 6,13-bis(triisopropylsilylethynyl)pentacene have been reported previously.[31]

Film Preparation

Thin films (ca. 90 nm) of each pentacene derivative were deposited atop glass slides via spin-coating at 2000 rpm from a 10 mg/mL chloroform solution. The glass slides were sequentially sonicated in water, acetone, and isopropanol prior to film deposition. Films that were amorphous following film deposition were subsequently crystallized via solvent-vapor annealing.

Femtosecond Transient Absorption Spectroscopy

The femtosecond transient absorption spectrometer has been described in detail previously.[30] Briefly, measurements were performed with a Ti:sapphire-based regeneratively amplified laser system (Coherent Libra, Santa Clara, CA) that delivers ∼45 fs pulses at a wavelength of ∼800 nm, a repetition rate of 1 kHz, and an average power of ∼4 W. Pump and probe beam paths were generated via a beamsplitter located at the output of the laser amplifier. A large fraction of the power was directed toward an optical parametric amplifier (Coherent OPerA Solo, Santa Clara, CA) that was used to convert the light to 645 nm, which was the pump wavelength used for the measurements. The pump and a small fraction of the power of the laser amplifier were directed toward a commercial transient absorption spectrometer (Ultrafast Systems Helios, Sarasota, FL). The latter was used to generate a continuum in the visible (ca. 420–760 nm) spectral region. An optical filter was used to isolate the continuum from the 800 nm light. The relative pump and probe polarization was controlled with a combination of a λ/2 waveplate and polarizer in the probe beam path located before the continuum generation crystal. Measurements were performed with pump and probe polarizations oriented at the magic angle (i.e., ∼55°) at the sample position. The pump beam spot size was determined by placing a digital CCD camera (Thorlabs, Newton, NJ) at the focal plane of the probe in the region of pump and probe overlap and analyzing an image obtained by using ThorCam software (Thorlabs, Newton, NJ). The spot size determined in this manner was ca. 180 μm. Pulse energies were measured with an optical power sensor and meter (Coherent, Santa Clara, CA). The incident pump fluence for the different measurements was typically in the amount of ca. 230 μJ cm–2, unless otherwise noted. For transient absorption measurements, films were secured in a filter holder situated on top of an automated linear translation stage. The films were raster scanned over the course of a measurement to mitigate the effects of photodegradation and increase overall scan time. The optical density of the films was generally kept below ca. 0.15 at the excitation wavelength.

Results and Discussion

We chose to study pentacene derivatives because the energetics of singlet fission are favorable, i.e., the process is exoergic, and because the molecular packing adopted by pentacene derivatives generally leads to electronic couplings favorable for triplet pair formation and separation. For these reasons, both triplet pair and subsequent triplet exciton yields are generally high, and each step of the singlet fission process can be directly time-resolved.[6−8,29,30] Here, we study several trialkylsilylethynyl (TAS)-substituted pentacenes, whose alkyl substituents directly impact their packing arrangements in the solid state. Figure displays the chemical structures and single-crystal molecular pair packing arrangements of two exemplary compounds: 5,14-TIPS-Pn and the prototypical 6,13-bis(triisopropylsilylethynyl)pentacene (6,13-TIPS-Pn). In the present work, we report on the singlet fission dynamics of thin films of 5,14-TIPS-Pn, molecule 1, and compare them with the complementary 6,13-substituted derivative, molecule 2. While triplet yields in solids of most pentacene derivatives studied to date, such as molecule 2, are quantitative, i.e., ΦT1 ∼ 200%,[6,7] the excited-state dynamics of molecule 1 have yet to be reported.

Figure 1

(a) Chemical structures, (b) representative local molecular packing, and (c) primary triplet pair decay pathway and typical triplet exciton yields for crystalline solids of 5,14- and 6,13-bis(triisopropylsilylethynyl)pentacenes. The crystal structure files of the former and latter, from which the depictions in panel b were obtained, correspond to Cambridge Crystallographic Data Centre (CCDC) numbers 172477 and 172476, respectively (see e.g. ref (31)).

As substitution of the pentacene core can profoundly influence both singlet and triplet state energies (see, e.g., section S3 of the Supporting Information), we first sought to determine if molecule 1 satisfies the energetic conditions necessary for singlet fission by studying amorphous films of molecule 1. As demonstrated in the pioneering work of Roberts and co-workers, amorphous films are, perhaps somewhat counterintuitively, a convenient material form to study singlet fission as the process can occur efficiently at dimer pair sites conducive to singlet fission.[5] Roberts et al., for example, showed that singlet fission takes place in amorphous 5,14-diphenyltetracene films in two stages: (i) an initial stage where dimer pair sites, which represent a subset of sites in the film, are directly photoexcited and (ii) a subsequent stage where primary (light-absorbing) singlet excitons migrate to these dimer pair sites. Roberts et al. further undertook a quantitative analysis of the triplet yields in the amorphous films and showed that singlet fission can be highly efficient. Highly efficient singlet fission has also been reported in amorphous solids of molecule 2 with triplet pair yields estimated at ΦTT > 99%.[6−8,27,29] Additionally, molecular pairs are not constrained to the molecular packing of equilibrium crystalline material, and there are generally a number of distinct slip- and twist-stacked configurations conducive to singlet fission.[33−36] Thus, we can determine if the energetic conditions necessary for efficient singlet fission (i.e., ES1 ≥ 2ET1) are satisfied in the absence of a preferred local packing. Amorphous films of molecule 1 (where due to the lack of any observable signal in grazing-incidence wide-angle X-ray scattering [GIWAXS] measurements, molecules are expected to exhibit at most only nearest-neighbor ordering[37,38]) were prepared via spin-coating; section S4 of the Supporting Information includes optical and structural characterization.

Figure a displays the transient visible absorption of the amorphous films of molecule 1; transient absorption spectra at selected time delays are presented in Figure b. At early time delays, i.e., ca. 100 fs, we observe a transient absorption spectrum consistent with primary singlet excitons. Specifically, ground-state bleach and stimulated emission features are observed in the vicinity of ca. 650 and 730 nm, and an intense photoinduced absorption band is observed at short wavelengths with amplitude at ca. 470 and 510 nm, similar to the singlet excitations of isolated chromophores (section S5). We therefore assign the early time transient absorption spectrum to primary singlet excitons. At longer time delays, such as at 400 ps, the transient spectra begin to resemble that of independent triplet excitations (section S6); we thus assign the long-time spectra to a population with considerable triplet character. Given that intersystem crossing is negligible on this time scale, as evidenced by transient absorption measurements on dilute solutions of molecule 1 (section S5), we assign the singlet-to-triplet conversion observed in the amorphous films to singlet fission. As described above, these triplet excitations form at a subset of dimer pair sites conducive for singlet fission in the amorphous films.[5,7,29] Consequently, we conclude that the energetic requirement for singlet fission must be met.

Figure 2

Transient visible absorption of amorphous films of 5,14-TIPS-Pn. (a) Surface plot of transient visible absorption. The experiments were performed with an incident pump wavelength of 643 nm and a fluence of 230 μJ cm–2. The scale bar is indicated beside the plot. The time axis of the bottom panel is plotted in linear increments, while that of the top panel is plotted in logarithmic increments. Ground-state bleach (GSB), stimulated emission (SE), and photoinduced absorption (PIA) features associated with the primary singlet excitons (i.e., S1S0) and triplet pairs (i.e., TT) are indicated beside the corresponding transient spectral features. (b) Selected transient spectra. At least three individual spectra were used to produce the time-averaged spectra. Spectra time delays are indicated in the legend. (c) Selected transient kinetics. The kinetics of the overlapping singlet and triplet photoinduced absorption bands and overlapping ground-state bleach and stimulated emission features were taken as the average over the spectral range 541–589 nm and 636–669 nm, respectively. These data have been normalized over the time range from 0.8 to 1.2 ps and 0.2 to 1.0 ps, respectively. A fit from a global target analysis overlays these transient absorption data (see section S7 of the Supporting Information for additional details).

Having demonstrated that molecule 1 satisfies the energetic requirements necessary for singlet fission, we proceeded to assay triplet yields in the amorphous films, an essential criterion for triplet harvesting. We previously showed that molecule 2, the related 6,13-substituted compound (Figure ), exhibits unity triplet pair yields, i.e., ΦTT, in the amorphous phase[6,29] and, on longer time scales relevant to triplet harvesting, independent triplet yields i.e., ΦT1, between 132% and 154%.[7,27]Figure c shows that while triplet pair yields in amorphous films of molecule 1 are also unity, significantly lower independent triplet yields (evidenced by the drastically reduced ground-state bleach amplitude on the few hundred picoseconds time scale) are observed. The analysis estimates quantitative triplet pair yields (>99.9%) based on kinetic competition between singlet fission and competing decay processes, such as unimolecular decay; specifically, singlet fission is assumed to be the primary nonradiative decay pathway in the amorphous films and the unimolecular lifetime and time constant for triplet pair formation are taken to be ≫500 and 0.6 ps, respectively (sections S5 and S7). In addition, the analysis uses the spectrally averaged amplitude of the ground-state bleach feature, normalized to the earliest time delays immediately following triplet pair formation, to quantify independent triplet yields of ∼48% (i.e., twice the relative amplitude at ∼400 ps). Although several factors complicate an accurate analysis of the triplet yield, as described in section S7, the results presented in Figure c suggest that differences in the local molecular packing between 5,14-TAS and 6,13-TAS substitution could be responsible for the discrepancy in the efficiency with which independent triplets are generated in singlet fission.

To induce long-range order, amorphous films of molecule 1 were crystallized by solvent-vapor annealing according to a prior report;[28] the optical and structural characterization of the films, which are polycrystalline with a crystallite size of ∼27 nm (calculated via the Scherrer equation) and that we hereafter refer to as “crystalline” films, can be found in section S4. Figure a displays the transient absorption of one such crystalline film, and selected transient spectra are presented in Figure b. At early times, near the region of pulse overlap, we observe a transient spectrum with features consistent with primary, light-absorbing singlet excitons. Namely, we observe a stimulated emission band in the vicinity of ca. 720 nm and a photoinduced absorption peaking at short wavelengths at ca. 510 nm. Figure b also displays two intermediate time delay spectra at ca. 300 fs and 1 ps. We assign these two transient spectra to two different types of triplet pairs and attribute the transition from the first to second triplet pair to intermolecular structural relaxation.[7,39] More details for the assignment of the excited-state dynamics can be found in section S8. Given that the area associated with ground-state bleach changes negligibly over the course of these dynamics up to ∼1 ps (Figure c) and the rate of conversion between the primary singlet exciton and triplet pair is much larger than that of unimolecular decay (sections S5 and S8), we conclude that essentially every primary singlet exciton is converted into a triplet pair in crystalline films of 1, i.e., ΦTT ≈ 100%.

Figure 3

Transient visible absorption of crystalline films of 5,14-TIPS-Pn. (a) Surface plot of transient visible absorption. The measurements were performed with incident pump wavelength of 643 nm and incident pump fluence of 230 μJ cm–2. The scale bar is indicated beside the plot. The time axis of the bottom panel is plotted in linear increments, while that of the top panel is plotted in logarithmic increments. The spectral region at negative and long-time delays where pump scatter is particularly noticeable has been highlighted in gray. GSB, SE, and PIA features associated with the primary singlet excitons (i.e., S1S0) and triplet pairs (i.e., TT) are indicated beside the corresponding transient spectral features. (b) Selected time-averaged transient spectra. At least three individual spectra were used to produce the time-averaged spectra. Spectra time delays are indicated in the legend. (c) Selected transient kinetics. The kinetics of the triplet pair photoinduced absorption and ground-state bleach features were taken as the average over the spectral range 500–560 nm and 560–710 nm, respectively. These data have been normalized over the time range from 210 to 440 fs and 500 to 650 fs, respectively. A fit from a global analysis overlays these transient absorption data (see section S8 for additional details).

Intriguingly, at longer time scales (400 ps), the photoinduced absorption band associated with the triplet pairs has decayed completely. Figure c shows that the decay of the photoinduced absorption band occurs with a time constant of ca. 2 ps. This decay is also reflected in the ground-state bleach feature (Figure c), where a derivative line shape appears at long times (Figure b). We attribute the derivative line shape to a thermal difference spectrum. This assignment is made on the basis that (i) the photoinduced absorption band, which describes transitions of an excited electronic population, disappears completely together with the ground-state bleach, and (ii) thermal effects are known to give rise to derivative line shapes in the vicinity of ground-state bleach features.[40,41] Our assignment is consistent with the conversion of a substantial amount of electronic energy, i.e., ca. 1.9 eV, into the nuclear kinetic energy degrees of freedom of system[42] and the subsequent dissipation of this energy into the bath. The dissipation of such a large amount of energy, which occurs either when one exciton decays directly to the ground state[7] or when two excitons annihilate,[40,41] raises the temperature of the medium locally and gives rise to a thermochromic effect. We can rule out both photodimerization and photo-oxidation as contributing to these decays on the basis that (i) both processes will disrupt conjugation of the pentacene core and lead to permanent photobleaching, which we do not observe, and (ii) the quantum yield for the sum of these processes is beyond the limit of detection of the transient absorption spectrometer (see, e.g., section S9). Thus, essentially every triplet pair decays directly to the ground state in crystalline films of 1, i.e., TT → S0S0 and ΦT1 ∼ 0%.

These results are in stark contrast to crystalline solids (i.e., films and nanoparticles) of 2, which have been characterized extensively. The consensus of these studies is that (i) primary singlet excitons are initially converted to triplet pairs within ∼100 fs[6,16,27,29] and (ii) these triplet pairs then dissociate via triplet pair separation on the few picosecond time scale.[30,43] Given the rapid nature of the first step in comparison with the ∼13 ns lifetime of monomer 2, the formation of triplet pairs is thought to occur with 99.999% efficiency; i.e., it is considered to be essentially lossless. Given that no changes to the amplitude of the ground-state bleach take place thereafter, the next step is also thought to take place without any losses. Thus, it is generally understood that triplet pair yields are unity, i.e., ΦTT ∼ 100%, in crystalline TIPS-pentacene, which ultimately translates into independent triplet exciton yields of ∼200%.[6]

We attribute the quantitative losses observed in crystalline films of 1, compared with the quantitative triplet yields observed in crystalline films of 2, to gross differences in local (pairwise) packing (Figure ). While the short-axis slip (or displacement, Δy in Figure b) between neighboring molecules is similar in crystals of 1 and 2, we find the long-axis slip (Δx in Figure b) to be drastically different.[31] Because both compounds exhibit small but comparable slip along their short axis that is associated with the transition dipole moment between the ground and first excited state, both compounds can undergo singlet fission.[1,2] Given the profound role side chains play in the packing of acene derivatives,[31,32,44] we attribute the difference in the slip between molecules along their long axis to side-chain sterics. Whereas the 6,13-substitution on molecule 2 prevents neighboring molecules from aligning in a center-to-center fashion, the 5,14-substitution on molecule 1 tends to promote center-to-center packing, giving rise to maximal π–π overlap.[45] Extensive π–π overlap has been shown previously to promote both excimer relaxation and rapid nonradiative decay of triplet pairs, via strong coupling between TT and S0S0, in covalently linked tetracene dimers.[46]

To further evaluate the impact of the 5,14-substitution pattern on excited-state dynamics, we investigated crystalline films of two additional 5,14-bis(trialkylsilylethynyl)pentacenes, molecules 3 and 4, and compared their excited-state dynamics with those of crystalline films of molecules 5 and 6, their 6,13-substituted variants (Figure ). Here, we hypothesize that triplet pair quenching is not inherent in the chemical nature of the substitution; rather, it is the molecular packing that is precipitated by the placement of the substitution that is responsible.

Figure 4

Chemical structures of two additional 5,14- and 6,13-bis(trialkylsilylethynyl)-substituted pentacenes, where the alkyl substituents are either ethyl or isobutyl groups.

Additional optical and structural characterization of crystalline films of molecule 3, 5,14-bis(triethylsilylethynyl)pentacene (5,14-TES-Pn), can be found in section S10. We previously showed that quantitative triplet yields via singlet fission are observed in crystalline films of molecule 5,[7] which adopts a similar local packing as molecule 2 (and note that this is the case even though their long-range order is different, i.e., molecules 5 and 2 adopt 1D and 2D packing arrangements, respectively[31,32,47]). Figure a,c,e displays transient absorption surface plot, selected spectra, and selected kinetics for crystalline films of 3. Similar to crystalline films of 1, a photoinduced absorption band in the vicinity of ∼500 nm and a stimulated emission band in the vicinity of ∼730 nm are observed at the earliest time delay, which decay rapidly. The rapid decay of these features occurs concomitant with the rise of a prominent photoinduced absorption band peaking at ∼550 nm. As for crystalline films of 1, we assign these excited-state dynamics to the decay of primary singlet excitons into triplet pairs. In contrast to crystalline films of 1, where the molecules adopt a preferential orientation with respect to the substrate (section S4) and the TT photoinduced absorption band is suppressed in amplitude relative to the main ground-state bleach feature (compare, e.g., Figure b and Figure S7.1b), in the case of crystalline films of 3 the TT photoinduced absorption band is much larger in amplitude compared with the main ground-state bleach feature. The latter result is consistent with the random out-of-plane orientation determined via GIWAXS (section S10). Specifically, whereas in the case of crystalline films of 1 the long axis of the molecules orient normal to the substrate, which is similar to unsubstituted pentacene and an effect known to suppress the intensity of the triplet photoinduced absorption relative to the main ground-state bleach feature,[48,49] in the case of crystalline films of 3 (and aqueous nanoparticle suspensions of 1) this effect is not observed. Crystalline films of molecule 3 exhibit complex kinetics, and so we have only shown the kinetics for a spectral region associated with overlapping S1S0 and TT photoinduced absorption bands (Figure e). Critically, these data indicate that all triplet pairs internally convert to the ground state with a time constant of ca. 2 ps.

Figure 5

Transient visible absorption of crystalline films of 5,14-TES- and 5,14-TIBS-Pn. (a, b) Surface plots of transient visible absorption. The measurements were performed with an incident pump wavelength of 650 and 664 nm, respectively. Both measurements were performed with an incident pump fluence of 230 μJ cm–2. The scale bars are indicated beside each plot. The time axes of the bottom panels are plotted in linear increments, while those of the top panels are plotted in logarithmic increments. GSB, SE, and PIA features associated with the primary singlet excitons (i.e., S1S0) and triplet pairs (i.e., TT) are indicated beside the corresponding transient spectral features. (c, d) Selected time-averaged transient spectra. At least three individual spectra were used to produce the time-averaged spectra. Spectra time delays are indicated in each legend. (e, f) Selected transient kinetics. For the crystalline 5,14-TES-Pn films, the kinetics of the combined S1S0 and TT photoinduced absorption is shown and was taken as the average over the spectral range 430–640 nm. These data have been normalized over the time range from 130 to 230 fs. For the crystalline 5,14-TIBS-Pn films, the kinetics of the ground-state bleach feature is shown and was taken as the average over the spectral range from 640 to 720 nm. These data have been normalized over the time range from 150 to 520 fs. A fit from a global analysis overlays these transient absorption data (see sections S11 and S13 for additional details).

We additionally investigated crystalline films of a 5,14-triisobutylsilyl (TIBS)-substituted variant, i.e., molecule 4. Additional optical and structural characterization of crystalline films of molecule 4 can be found in section S12. Figure b,d,f displays transient absorption surface plot, selected spectra, and selected kinetics for crystalline films of 4. As was the case for the other 5,14-substituted pentacenes, spectral features consistent with primary singlet excitons are apparent at the earliest time delays, these features decay, and then the entire transient signal decays quite rapidly. There is no obvious TT photoinduced absorption band, which is likely related to the preferential out-of-plane orientation of the molecules identified via GIWAXS. As noted above, such out-of-plane orientation may suppress the amplitude of the TT photoinduced absorption band compared to the main ground-state bleach feature. As we saw when comparing crystalline films of 2 and 1 above, and crystalline films of 5 and 3, we again find that while in the case of 6,13 substitution (i.e., molecule 6) triplet pair yields are high and triplet excitations are long-lived,[29,30] in the case of 5,14 substitution (i.e., molecule 4) efficient internal conversion to the ground state and complete triplet pair quenching are observed (Figure b,d,f and section S13).

On the basis of the results above, we arrive at the following picture (Figure ). While the local molecular packing (and long-range order) adopted by 6,13-TAS-Pn seems to be generally favorable for singlet fission,[50] i.e., triplet pair dissociation is efficient and triplet yields are high, the molecular packing adopted by 5,14-TAS-Pn seems to be generally unfavorable for singlet fission, i.e., direct internal conversion to the ground state is the primary decay pathway and triplet pairs are completely quenched. We recently identified an H-aggregate molecular packing unfavorable for singlet fission in amorphous solids of 6,13-TAS-Pn derivatives[7] whose properties are generally consistent with those reported in the present work for crystalline films of 5,14-TIPS-Pn. In that work, we also discussed the nature of the electronic structure of these triplet pairs and their decay mechanism. Here, we further posit that the nonradiative decay is promoted by strong nonadiabatic coupling between TT and S0S0. We note that intermolecular structural relaxation, i.e., “excimer” relaxation,[39,51−53] also appears to play a role in the photophysics of these 5,14-TAS-Pn.

Figure 6

Local (pairwise) packing and state diagram describing singlet fission in crystalline solids of 5,14-TIPS- and 6,13-TIPS-Pn. (upper panel) Local packing derived from the single crystal structure for 5,14-TIPS- and 6,13-TIPS-Pn, illustrating how 5,14-TIPS substitution promotes very little long-axis displacement whereas 6,13-TIPS substitution causes significant long-axis displacement. The crystal structure files of the former and latter, from which the depictions were obtained, correspond to CCDC numbers 172477 and 172476, respectively (see, e.g., ref (31)). (bottom panel) State diagrams depicting singlet fission in crystalline solids of 5,14-TIPS- and 6,13-TIPS-Pn. For 5,14-TIPS-Pn, the energy for the triplet pair was estimated as 2 × ET1, or 1.6 eV, where ET1 was taken to be 0.8 eV, based on the values of for the triplet energy of unsubstituted and 6,13-substituted TIPS-pentacene are 0.85 and 0.78 eV, respectively,[54,55] and given that conjugation with the 5,14-substitution is expected to decrease the triplet state energy with respect to unsubstituted pentacene. For 6,13-TIPS-Pn, the energy of the triplet pair was estimated as 2 × ET1, or 1.56 eV, where ET1 was taken to be 0.78 eV.[55] Black arrows represent transitions between states, large black arrows indicate the primary TT decay pathway, and gray arrows represent transitions not observed in the present work. We note that intermolecular structural relaxation, i.e., “excimer” relaxation, also appears to play a role in the excited-state dynamics of 5,14-TIPS-Pn. The range of time constants associated with internal conversion of triplet pairs (i.e., τIC) shown were obtained from the present work, while the range of time constants associated with triplet pair separation (i.e., τTPS) were obtained from past work.[29,30]

Similarly drastic nonradiative decay of triplet pairs has been observed in covalently linked pentacene and tetracene dimers. For example, Zirzlmeier et al. synthesized ortho-, meta-, and para-phenylene-linked ethynyl pentacene dimers and observed rapid and efficient triplet pair formation in all the dimers; in the ortho-substituted dimer, which exhibited the most π-orbital overlap, the triplet pair lifetime was only ∼12 ps.[55] In another work, Sanders et al. synthesized a series of 2,2′-bipentacenes bonded directly and separated by one and two phenylene spacer groups.[56] The authors found that as the distance between the pentacene molecules was reduced, i.e., as spacer groups were removed, both the time scales of triplet pair formation and decay also decreased, with the shortest triplet pair lifetime measured at ∼450 ps. A similar relationship was recently observed by Paul and Karunakaran in bipentacenes covalently linked at their 2,2′-, 2,6′-, and 6,6′-positions with an ethynyl bridge.[57] Sanders et al. also observed that the bipentacene with the longest spacer exhibited a triplet pair lifetime of ∼270 ns. These results were consistent with Lukman et al., who synthesized an orthogonal pentacene dimer and observed a similarly short triplet pair lifetime of ∼650 ps, which could be extended in certain polymer matrices to as long as ∼140 ns.[58] The benchmark in all of these cases, to which the triplet pair lifetime is to be compared, is the ∼10 μs lifetime of the isolated triplet excitation of 6,13-TIPS-pentacene.[7,55,56,59] Lastly, Korovina et al. synthesized two ethynyl-substituted tetracene dimers with distinct phenyl and xanthene bridging groups.[46] The authors found that both tetracene dimers exhibited short triplet pair lifetimes of the order of hundreds of picoseconds and, consistent with Zirzlmeier et al.,[55] that the tetracene dimer with the most π-orbital overlap exhibited the shortest triplet pair lifetime of ∼100 ps. Critically, the authors also performed theoretical calculations indicating that excimer relaxation takes place in the covalently linked tetracene dimers and that strong coupling between TT and S0S0, particularly in the excimer geometry, mediates the nonradiative decay of the triplet pairs. While systematic studies of covalently linked dimers, such as those highlighted above, have provided considerable insight into singlet fission dynamics, including triplet pair quenching (with additional insights provided in the most recent reports[60,61]), it is hoped that the distinct intermolecular systems reported in the present work may add to this body of literature and aid in shedding additional insight into the origin of triplet pair quenching in singlet fission materials.

Conclusions

In summary, we reported a local (pairwise) packing arrangement in crystalline films of 5,14-bis(triisopropylsilylethynyl)-substituted pentacene that is detrimental to singlet fission. We showed that while triplet pairs form in quantitative yields, they decay rapidly via strong coupling to the ground state. Thus, quantitative triplet losses are observed. We showed that quantitative losses are observed in crystalline films of two additional 5,14-bis(trialkylsilylethynyl)-substituted pentacenes with either ethyl or isobutyl substituents. Thus, avoiding the packing arrangement in these 5,14-substituted pentacenes is paramount and motivates an investigation into the influence of similar packing arrangements on the dynamics of other singlet fission materials.