2.1. Excited State and Structural Dynamics in Oligomers
In order to discuss the ultrafast dynamics below, structures of the PTB7
oligomers are shown in Figure 1
a, along with their steady state absorption and fluorescence spectra in Figure 1
b–c, respectively. As demonstrated previously [20
], the length-dependent absorption maxima red-shifts as n increases, but the shift saturates at n > 3 due to conformational disorder concerning the C-C bonds connecting adjacent repeating units, which attenuates the π-conjugation along the backbone [27
]. It is worth noting, however, the fluorescence spectra in Figure 1
c do not strictly follow this pattern and further shifts could happen in oligomers with n > 3 following the observed trend. Such differences in the trend of red-shift may reflect conformational differences between the absorbing and emission species for these oligomers in terms of π-conjugation lengths for the ground and excited species.
To further investigate, we performed time-resolved fluorescence measurements using a streak camera and the results are shown in Figure 2
. The n = 1 oligomer was unstable under the laser excitation and its results are shown in the SI (Figures S16 and S17, Tables S6 and S7)
. The time-resolved fluorescence spectra for n = 2 and n = 3 oligomers are shown respectively in Figure 2
a–c. Both contain fast spectral evolution on the blue edge of the spectrum and longer-lived fluorescence at the red edge of the spectrum, which are represented in the two-component decay-associated spectrum in Figure 2
b,d (see Methods and SI
for fitting details) from the global fitting of the data. Beginning with the n = 2 oligomer in Figure 2
b, the fit contains two components: 1) a species (blue) with short decay time constant that largely constitutes emission centered at 580 nm and 2) a species (orange) with a decay time constant of 1200 ps that constitutes broad, red emission from 600 to 700 nm. When oligomer length is increased from n = 2 to n = 3, the two components are largely retained with key differences only in the peak position and decay time constants. Each component is red-shifted, most notably the longer-lived (orange) peak, which shifts 1144 cm-1 from a maximum at 618 to 665 nm.
This shift in the long-lived species can begin to explain the discrepancy between the peak maxima trends in absorption and fluorescence spectra discussed above. It has been widely reported that exciton formation in conjugated polymers tends to trigger transformation of the conjugated backbone from a benzoidal to a quinoidal form, resulting in a planarized backbone [28
]. In fact, this change was a guiding principle in the design of PTB7 [30,31,32]
that appears to be retained in the oligomer series. This planar conformation of the excited state potential energy minimum allows for increasing π-conjugation in the excited state that overcomes the conformational disorder inherent to the ground state and leads to the continuing red-shift with increasing n seen exclusively in the fluorescence spectra. These long-lived, red-shifted (orange in decay-associated plots) spectral features can then be attributed to the emission from the planar, quinoidal excited state. The magnitude of the red-shift of the steady state fluorescence maxima with increasing n will involve a natural red-shift of this emitting feature and the relative intensity of this quinoidal state relative to the blue emitting species. Because the overall photophysical behavior is otherwise retained, the shifting of the emission maxima of this species appears to be the primary effect of chain lengthening in the absence of aggregation.
Experiments with higher time resolution were required to identify the blue fluorescent species and provide further evidence for the identification of the quinoidal excited state. Towards this end, transient absorption (TA) and a higher time resolution streak camera experiment were conducted with specific emphasis on the n = 2 oligomer for two reasons: 1) besides peak shifting and small lifetime differences, ultrafast dynamics are similar between n = 2 and n = 3 oligomers, and 2) the n = 2 oligomer has its emission spectrum aligned with that of the corresponding polymer, and hence, is a useful model for unfolded portions in PTB7
]. The parallel results for n = 3 oligomer are available in the SI (Figures S18–S22, Tables S8–S12)
Complementary to the time-resolved fluorescence measurements, we also obtained ultrafast transient absorption spectra of the n = 2 oligomer in the visible region at 550 nm excitation. It features both ground state bleaching (GSB) from 500–600 nm and stimulated emission (SE)/excited state absorption (ESA) from 550 to 750 nm as shown in Figure 3
b, mirroring their respective steady state features above in Figure 3
a. There is a clear distinction in time evolution of the spectral features in the region around 600 nm with much faster decay than the rest of the spectrum, leaving behind spectrally distinct bleaching and emission peaks at later times. These data were fit globally, yielding the two-component decay-associated spectrum in Figure 3
c (see SI Figure, Table S1
for fitting details).
The decay-associated fitting again reveals two characteristic spectra with two respective decay time constants, 14 and 1000 ps. The interpretation of transient absorption spectra is more complex due to the overlapping GSB, SE, and ESA spectral features. The ~1000 ps orange feature constitutes a majority of the ground state bleaching signal, overlapping nicely with the ground state absorption spectrum and identifying this pathway as the main avenue for relaxation to the ground state. It also has a lifetime and broad emission component from 600 to 700 nm similar to the red-shifted emission (Figure 2
b) attributed to the planar, quinoidal excited state. Meanwhile, the blue decay component in Figure 3
c largely constitutes emission at 600 nm, spectrally similar to the blue-shifted fluorescence component (Figure 2
b), which can now confidently be assigned a lifetime of 14 ps. From 500 to 600 nm, this blue component has a small contribution to the ground state bleach feature, indicating some minority relaxation directly to the ground state. Evidence for the majority relaxation pathway of this species can be seen in the positive feature between 650–700 nm. Because the transient signal is negative, this positive feature indicates a growth in that region during that 14 ps time component. Evidence for this growth exists at 675 nm in Figure 1
b, where the signal increases in magnitude between the presented 2 and 25 ps time cuts. Taken as a whole, it seems the 14 ps, blue-shifted species primarily decays into the 1000 ps, red-shifted species.
These observations are best understood in the context of excited state dynamics in an ensemble of conformational isomers with varying degrees of planarity. As previously discussed, the oligomer ground state allows for inhomogeneity in dihedral angle and π-conjugation between adjacent benzodithiophene (BDT) and thienothiophene (TT) units, while the quinoidal character of the excited state moves the energetic minimum towards more planar conformations. As such, the excitation pump pulse will initially excite an ensemble of oligomers with varying dihedral angles at various distances from their new, planar, excited state energetic minimum. The shorter lifetime, blue-shifted emission species are composed of those oligomers that are more severely twisted upon excitation, and their 14 ps lifetime is not a “fluorescence lifetime,” which is typically much longer for conjugated organic chromophores [33
], but is instead their torsional relaxation time, in line with timescales seen in thiophene-based oligomers [28
]. As evidenced by their presence in fluorescence spectra, there is some emission from these species during their relaxation period, but a majority torsionally relax into the ~1-ns, quinoidal, planar excited state. The longer-lived, red emission, then, is the primary relaxation pathway of the oligomer and is composed of both those oligomers that were relatively planar upon initial excitation and those that underwent dynamic planarization.
To directly observe this conversion from twisted to planar excited species, we carried out a second streak camera measurement with the overall time window of just ~100 ps, with ten times enhanced time resolution compared to the first set of measurements (see details in Methods section). This set of measurements enabled us to observe kinetics for the blue-shifted, twisted species while compromising the measurements for the longer component, which had a lifetime longer than the observation window. Figure 3
d shows traces of the far blue and far red edges of the fluorescence spectrum of the n = 2 oligomer and their fit lines, chosen despite their comparatively low signal to noise ratio because they contain emission exclusively from the twisted and planar species, respectively. The growth of both regions begins concurrently but diverges after a few ps. While the blue emitting species reaches its maximum and begins to decay within the instrument response time, the redder species shows a secondary growth, not reaching its maximum value until the blue species has decayed and providing direct evidence for interconversion over the course of 10–20 ps. Fluorescence upconversion scans, which have much better time resolution and are available in the SI (section S1d)
, largely agree with kinetics observed in the streak camera and provide additional evidence of dynamic planarization.
2.2. In-Chain Effects on CT Character.
s intrinsic backbone charge transfer character has been credited with many of its unique properties implicated in its high efficiency exciton splitting and photovoltaic device performance [24
]. As such, the exact nature of this charge transfer character has been of great scientific interest. However, studies have been hampered by one simple fact: PTB7
is notoriously insoluble in all but a few chlorinated solvents [9
], making it difficult to observe solvent-dependent emission shifts that reflect charge transfer characteristics [22
]. The PTB7
oligomer series, then, allows for a unique opportunity to study charge transfer character in the PTB7
exciton for two reasons: (1) they retain solubility in a variety of solvents and (2) the geometry of their flattened, quinoidal excited state resembles the self-folded PTB7
exciton. To obtain an accurate comparison with the polymer, fluorescence streak camera measurements were carried out with a time window extending to 1200 ps and a pump wavelength of 500 nm in solvents of varying polarity, and the contribution of the longer-lived, quinoidal excited state was isolated by integrating and normalizing intensities from 200 ps onwards, as shown in Figure 4
a. The most notable shift occurs on the red edge of the fluorescence spectrum, where the polar solvents have their edge shifted ~500 cm−1
(20% peak height at 705 nm in toluene vs. 731 nm in chloroform/THF), indicating an increase in dipole moment of the quinoidal excited state relative to the ground state [36
] and substantial charge transfer character. This indicates that PTB7′
s polar exciton is inherent to its “push–pull” backbone structure.
2.3. Aggregation Effects on CT Character.
Beyond a polar exciton, a defining feature of PTB7′
s near-infrared probe ultrafast spectra is its characteristic polymer cation peak [22
], in which its “charge transfer” and “charge separated” states were detected and described as excitons that, almost instantaneously after formation, split into holes and electrons with varying extents of separation. Importantly, these spectral features were identified by comparisons to polymer cations observed in both bulk heterojunction films and those obtained via spectro-electrochemical measurements [22
]. Because analogous charge generation has also been seen in neat conjugated homopolymers in film (but, to our knowledge, not in solution) without this “push–pull” structure [39
], it is reasonable to ask if this ultrafast charge generation is a property inherent to PTB7′
s backbone or if it is an emergent property of the self-folded polymer chain. If ultrafast charge separation occurs in the oligomers akin to that seen in the polymer, there should appear transient signal of charged oligomer, especially the cation which is known to appear in the spectrally accessible near-infrared region [38
To obtain the spectra of the oligomer cation, spectro-electrochemical measurements were performed, as shown in Figure 4
b. Two changes were observed: blue-shifting of the absorption maximum and small induced absorption at ~900 nm (Figure 4
b, inset). Importantly, this blue-shifted absorption was not observed in visible TA experiments as detailed above (Figure 3
b), providing evidence against the existence of charged oligomer species.
For further comparison to the ~900 nm cation peak (Figure 4
b, inset), ultrafast TA experiments were conducted on the n = 2 oligomer with a near-infrared probe, as shown in Figure 4
c, with decay-associated fitting in Figure 4
d. There are two broad features in the TA spectrum in Figure 4
c, centered blue of 875 nm and at 1100 nm. Noting that the time constants for these species are comparable to those seen in the visible region discussed earlier, these excited state absorption features are assigned to the twisted (875 nm) and planar (1100 nm) excited species. Further confirmation of the 1100 nm peak as the planar excited species is seen in its spectral red-shifting during the first ~20 ps with the dynamic planarization process, as seen clearly in Figure 4
c and represented by the positive feature of the 8 ps component in the decay-associated spectrum in Figure 4
d at ~1000 nm. Because neither excited state absorption signature aligns with the near-infrared spectrum obtained for the oligomer cation, the near IR probe TA experiment indicates no photoinduced formation of oligomer cation species. In fact, the ~1100 nm excited state absorption of the n = 2 oligomer observed in the near-infrared is consistent with linear response calculations of the excited singlet absorption of a short oligomer of PTB7
] indicating that the planar, quinoidal species is a singlet excited state and that the progression from charge transfer character to charge separation is modulated by the self-folded polymer structure. Importantly, the appearance of charge separation only in the self-folded polymer does not assert that exciton splitting necessarily occurs only across multiple chains, as the enhanced π-conjugation in the polymer could allow further modulation of the charge transfer character of the backbone beyond that seen in even the n = 2 and n = 3 oligomers.
The identity and properties of the ~1100 nm excited state absorption signature in neat PTB7
is still an area of active research, due to overlapping spectral signatures of the polymer triplet absorption [41
] with our assignment of the polymer cation [22
] and recent work suggesting the ultrafast creation of symmetry conserving triplet pairs in PTB7
]. While the presence of triplets at longer times in PTB7
is uncontroversial [41
], we contend that the species produced in the ultrafast regime is a charge separated state despite the energetics of exciton splitting without electron acceptor at first seems energetically unfavorable: PTB7
has been estimated to have thermally inaccessible exciton binding energies of ~0.4 eV [42
] or as low as ~0.2 eV [43
] at low torsional angles. Nevertheless, measurements with externally applied fields did not observe the step-like transition predicted by a simple Onsager–Braun model with these kinds of splitting energies [44
] at the applied field value that overcomes the exciton binding energy. Instead, it was observed that exciton splitting shown large dependencies on disorder and a gradual onset with respect to the applied field at energies lower than those predicted by an Onsager–Braun model. This suggests an inherent sensitivity to the polymer local environment and site energies [44
], as small fractions of polymer meet the energetic requirements for exciton splitting before the bulk film. In fact, this may suggest a mechanism by which aggregation modulates the CT character inherent to the PTB7
backbone discussed in this manuscript. Variations in π-conjugation length in PTB7
chains have shown very large energetic shifts via the oligomer series, with over 0.37 eV bandgap shifts between oligomer and polymer [20
]. Assuming symmetrical movement of HOMO and LUMO levels, this kind of shift could lead to relative movement of ~0.19 eV of polymer and oligomer frontier orbitals. The presence of solution aggregates provides both the energetic shifts and disorder in site energies necessary for exciton splitting. Nevertheless, these transient species, and any long-lived spectroscopic species, are not present in the oligomers, as shown in this work. Regardless of the mechanism, the modulation of energetics via aggregation and local disorder is necessary for whichever process dominates.