Laser-induced fluorescence spectroscopy of TIPS-pentacene Open Access

Polycyclic aromatic hydrocarbons (PAHs), particularly polyacenes, are emerging as strong contenders for use in organic semiconductors and optoelectronic devices. Acenes have already been utilized in various applications, including field-effect transistors,¹ organic sensors,² and as emissive layers in OLEDs.³ Modifying these molecules enables fine-tuning of their physical and electronic characteristics, such as the HOMO–LUMO gap, solubility, stability, and phase behavior, to fit requirements for specific applications.⁴ 

One highly promising strategy involves attaching triisopropylsilyl (TIPS) side groups to acenes, depicted in Fig. 1. This modification enhances the molecules’ physical stability and solubility⁵ and optimizes the packing arrangement in thin films^6,7 resulting in a brick wall structure compared to the herringbone structure of pentacene (Pc).^8,9 This results in enhanced π–π interactions leading to a more efficient charge carrier transport.^5,10

Moreover, acenes, particularly TIPS-pentacene, are recognized for their ability to undergo singlet fission, where the absorption of a single photon generates two triplet excitations.^11–14 This process holds the potential to surpass the Shockley–Queisser limit,^15,16 with quantum yields exceeding 200% already being recorded in TIPS-pentacene solutions.¹⁷ 

Many studies have investigated the electronic properties of TIPS-pentacene and the development of various thin films, often comparing it to the unmodified pentacene molecule and other functionalized derivatives.^18,19 Furthermore, solution-based experiments have shed light on the energetic structure, molecular dynamics, and interactions of individual molecules.^20,21 In this context, vibrational breathing modes were examined using 2D electronic spectroscopy techniques. Also, comparisons of the photoionization of TIPS-pentacene and pentacene were reported in the gas phase and thin films.^22,23

While not all details of the mechanisms behind singlet fission are fully understood, recent experimental studies applying transient absorption and 2D electronic spectroscopies on thin films of pentacene and some of its derivatives^13,14 have indicated that coherent vibrational motion involving several vibrational modes is involved in the process. The complex experimental procedures and partially overlapping bands make it challenging to unambiguously attribute the observed vibrational frequencies to certain electronic states. High-resolution Laser-induced fluorescence (LIF) spectroscopy can identify the vibrational modes coupled to the S₀–S₁ transition, that corresponds to the vibrational modes contributing to vibrational wave packets formed in this optically accessible electronic state upon excitation with broad-band laser pulses.

High-resolution LIF spectra of the unsubstituted pentacene are widely reported and discussed in previous studies either as a cold molecular beam²⁴ or utilizing helium nanodroplet isolation spectroscopy.^25–27 Furthermore, the spectral response of the pentacene molecule attached to solid rare-gas clusters like argon and neon was published and discussed.^28–31 

Similar high-resolution spectroscopy studies on TIPS-pentacene have however not been previously reported. We therefore present here a study comparing the high-resolution LIF excitation spectrum of unsubstituted pentacene to TIPS-pentacene isolated in superfluid helium nanodroplets and on solid rare gas clusters. This comparison will shed light on the differences of the vibronic structures of these molecules and the way they are influenced by the different inert matrices. Furthermore, detailed knowledge of the Franck–Condon active vibrational modes in the optically accessible states of TIPS-pentacene will help in the future analysis of vibrational coherences recently reported to contribute to ultrafast singlet fission processes.^13,14

The experimental setup for laser-induced fluorescence measurements employed in this study has been largely described in prior work.²⁸ Rare-gas clusters and helium nanodroplets are generated through supersonic expansion of gases at high pressures through a pulsed Even–Lavie valve^32,33 with a 60 μm nozzle orifice. The valve is cooled using a two-stage Gifford–McMahon cryocooler (Sumitomo, RDK-408D2) to achieve subcritical expansion and clustering conditions: He (p = 50 bar, T = 18 K), Ne (p = 30 bar, T = 90 K), and Ar (p = 50 bar, T = room temperature). Cluster sizes were estimated using scaling laws^33,34 and monomer doping curves, yielding approximately 450 atoms for argon, 12000 atoms for neon, and 30000 atoms per helium droplet. TIPS-pentacene (Ossila, purity > 99%) is introduced to the clusters or helium droplets via the pick-up technique,^35,36 ensuring monomer-level doping for each individual cluster or nanodroplet.

Subsequent excitation of the molecules is performed using a pulsed dye laser system (Sirah Cobra), pumped by a frequency-doubled or tripled Nd:YAG laser (Edgewave Innoslab IS8II-E or IS8III), providing a broad wavelength range. The excitation laser has a pulse duration of 3.2 ns (FWHM) and a beam profile of 2 mm² in the interaction region. Pulse energy ranges from 3 to 12 μJ, corresponding to power densities between 12 and 48 kW/cm². All spectra are normalized, with background signals, stray light, and fluorescence from effusive molecules subtracted.

LIF excitation spectra of TIPS-pentacene attached to rare-gas clusters or embedded in helium nanodroplets were recorded. The band origin of the transition is presented in Fig. 2, along with the corresponding spectra of pentacene under the same cluster or droplet conditions. To compare the excitation spectra, all pentacene spectra were red-shifted in Fig. 2 by approximately 1850 cm^–1, as indicated in the individual panels. The spectra were normalized to their maximum intensity, and the wavenumber axis covers a range of 250 cm^–1 in all three panels. For helium nanodroplets, a single dopant molecule is dissolved within the droplet, whereas in solid neon and argon clusters, the molecule is attached to the cluster surface.

Table I summarizes the positions of the band maxima and corresponding FWHM values for the electronic transition from the ground state with ¹A_g symmetry to the first electronically excited state with ¹B_2u symmetry.³⁷ The reported values for spectra on neon and argon clusters are determined by fitting Gaussian peaks, while the values for helium nanodroplets are the maximum of the band. The uncertainty of the band positions retrieved from the fitting procedure is 1 cm^–1. The measured peak positions and widths of the pentacene excitation spectrum are in agreement with previously reported data.^25,31,38 Notably, the maximum peak position of pentacene spectra on solid rare-gas clusters does not represent the pure electronic band origin; for example, for pentacene on neon clusters, a reported redshift of about 24 cm^–1 of emission spectra relative to excitation spectra,³¹ indicates vibrational contributions of the hindered motion of the molecules on the neon surface as well as coupled vibrations of the neon cluster.

The observed energy difference between TIPS-pentacene and pentacene excitations remains within a narrow range between 1827 and 1887cm^–1 across all three media. For the different acenes it has been previously reported,²⁸ that the excitation energies of the different molecules are shifted by the same amount depending on the choice of rare gas environment. This trend seems to extend also to TIPS-pentacene.

The spectral linewidths for molecules on solid clusters are generally determined by the properties of the clusters. Broadening is introduced by the coupling of the electronic excitation to low frequency vibrational modes of the clusters as well as intermolecular modes between the cluster and the molecule that cannot be completely frozen at the cluster temperatures of 10 K (neon) and 37 K (argon).³⁹ Further contributions to the linewidth come from the dissipation of the corresponding vibrational energy in the cluster, as well as from possible inhomogeneous effects on small clusters.⁴⁰ Consequently, the linewidths on argon clusters, with higher temperatures and stronger interactions between the molecules and the clusters are larger. The apparently narrower linewidth of TIPS-pentacene on argon compared to pentacene can be attributed to the first vibrational mode of pentacene, known as the out-of-plane “butterfly mode” at around 85 cm^–1.²⁴ While this mode is well-resolved in neon clusters, it appears as a shoulder in the argon cluster spectrum. This bending motion of the pentacene backbone along the long molecular axis²⁴ is significantly suppressed in TIPS-pentacene, reducing the apparent linewidth in the spectrum of TIPS-pentacene on argon. This difference between TIPS-pentacene and its unsubstituted counterpart indicates differences in the corresponding molecular potentials for example an increased stiffness of the TIPS-pentacene molecule caused by the slightly more extended π-system.

A significant difference in the line shape at the band origin is observed in the excitation spectra of both molecules embedded in helium nanodroplets, as shown in Fig. 2(a). The linewidth of pentacene is below 0.5 cm^–1 due to the non-interacting, ultra-cold helium environment (0.4 K). In helium nanodroplet isolation (HENDI) spectroscopy, organic molecules typically exhibit a narrow zero-phonon line at the electronic band origin followed by a separated phonon wing, a characteristic feature of the quantum nature of the helium environment.^25,41,42 Contributions of vibrational hot bands observable at frequencies below the electronic origin are suppressed even for the lowest-frequency vibrational modes.

In contrast, the TIPS-pentacene spectrum shows a broad band with a distinct but only partially resolved substructure. Furthermore, the distinct features in this substructure are significantly broader than the electronic origin line of pentacene in helium nanodroplets. Consequently, a clear electronic origin as the energetically lowest possible transition is not identifiable. The substructure may be attributed to low-energy vibrational modes and their progressions, with the electronic origin either at the red end of the broad structure or outside the Franck–Condon window. Organic molecules with large side groups, such as TIPS-pentacene, are known to have low-energy vibrational modes associated with bending or torsional modes^43–45 of the side groups. The absence of these vibrational progressions in the spectra of TIPS-pentacene on solid rare gas clusters points to differences in the interactions of the molecules with the rare gas clusters, effectively fixing the molecular side groups in space, or drastically damping their motion. Similar effects have been previously reported for out-of-plane bending modes of pentacene, that can be observed in excitation spectra of isolated molecules but not in the condensed phase.²⁴ 

Significant broadening of individual vibronic lines in superfluid helium nanodroplets has been reported in the past from two effects.^43,45 The first is lifetime-broadening resulting from effective damping of large-amplitude vibrational motions. Such damping effects are caused by effective coupling between low-energy molecular vibrations and the helium environment. This effect would, however, not affect the electronic origin band (0–0 line), which should instead remain a sharp feature and the lowest frequency transition in the spectrum. For broad vibrational progressions, caused by excited state equilibrium geometries differing from the ground state geometry, the 0-0 line may have very low intensity due to low Franck–Condon overlap. A second effect causing severe broadening in spectra of molecules isolated in helium nanodroplets, is the severe perturbation of the excited state electronic density by the helium environment.^43–45 Given the similarity of the pentacene and TIPS-pentacene molecules and the absence of such perturbations for pentacene this second possible explanation seems unlikely. We thus suggest a broad Franck–Condon progression of damped low-frequency vibrational modes as the origin of the broad, structured band.

Considering the molecular structure of TIPS-pentacene, the torsional modes originating from rotations around the C-Si bonds depicted in Fig. 1 appear as the most probable but not only candidates. While the pentacene backbone and alkyne groups are relatively rigid and stabilized by the π-electron system, rotation can occur at the C-Si bonds with a shallow torsional potential. A computational study reported equilibrium torsion angles of 87° and 68° (relative to the plane of the pentacene backbone, see Fig. 1) for the two TIPS groups in the electronic ground state.⁴⁶ The broken symmetry in the ground state suggests that a change in equilibrium angles in the electronically excited state is plausible.

Similar although much less pronounced differences between pentacene and TIPS-pentacene have been previously reported for photoionization spectra,²² characterizing the transition of the neutral molecules to their cationic ground state. While in the case of pentacene, the vibrational structure of the ionizing transition has been analyzed in some detail in terms of involved vibrational modes and relaxation energy,⁴⁷ less details have been reported for the case of TIPS-pentacene. However, the less-resolved nature of the TIPS-pentacene photoelectron spectrum indicates the possible additional contribution of unresolved low-frequency vibrational modes to the photoelectron spectrum and relaxation energy of the molecular cation.

Figure 3 shows the vibrational excitations of the first electronically excited state of TIPS-pentacene for different environments. For better comparison the origin of the abscissa has been fixed to the maximum intensity of the lowest-frequency signal, despite this maximum not corresponding to the pure electronic excitation and spectral intensities have been scaled as indicated for better visibility of the bands at higher vibrational excitation. For comparison we also show the spectrum over the corresponding range of excitation energies for unsubstituted pentacene in helium nanodroplets. Note that the latter spectrum may be slightly saturated and is not corrected for the frequency dependence of the laser power.

In the first approximation, the vibrational structure of TIPS-pentacene appears to be independent of the surrounding matrix. All vibronic excitations of TIPS-pentacene can be observed in all three rare gas environments at identical excitation energy relative to the lowest-frequency signal. The sole exception to this observation is the substructure of the lowest frequency signals in helium nanodroplets discussed already above. Upon close inspection, it is apparent that the shape of the band system close to the electronic origin with its substructure is reproduced in many of the individual vibrational modes. This is especially evident for the vibrational band around 250 cm^–1 in the TIPS-pentacene spectrum in helium droplets, which closely resembles the shape of the band system at 0 cm^–1, while it is less evident for the vibrational excitations above 1000 cm^–1 (see Appendix Fig. 4). It remains unclear if the substructure is not present for these high-frequency vibrational modes, or if it is just disguised by additional lifetime effects and slightly overlapping bands. The neon spectrum, with its minimal linewidth and symmetrical profile, provides the best conditions for identifying distinct vibrational modes. Nevertheless, most vibrational features are also discernible in the spectra of other media, provided that the linewidth of each peak is narrow enough to distinguish between two transitions.

Table II summarizes the identified vibrational transitions, their possible assignments, and comparisons with existing data on TIPS-pentacene and unsubstituted pentacene. Vibrational frequencies of the low-frequency modes of pentacene in helium droplets were obtained from LIF excitation spectra²⁵ and are reconfirmed by our own spectrum in Fig. 3. Vibrational data on TIPS-pentacene comes from Raman spectroscopy of powder and thin film samples, accessing the vibrational frequencies of the electronic ground state. Further experimental data on TIPS-pentacene was reported from 2D spectroscopy of a dilute TIPS-pentacene solution²⁰ and transient absorption¹³ as well as 2D electronic spectroscopy¹⁴ of thin films. In the low frequency range, both molecules show close correspondence in the observed vibrational modes of the first excited states. The reduced intensity of the out-of-plane butterfly mode in the case of TIPS-pentacene has been mentioned above. We can add to this that also the frequency of this mode is slightly changed in TIPS-pentacene. If the feature at 190 cm^–1 corresponds to the next peak in the progression of this band, this indicates a more anharmonic butterfly mode than in the case of unsubstituted pentacene. The in-plane ring breathing mode (256 cm^–1) of the pentacene moiety and the second peak in its progression (512 cm^–1) appear very similar in both molecules in terms of frequency and intensity, as can be expected for a mode that is largely localized on the pentacene backbone. An additional mode at 604 cm^–1 previously reported for pentacene emission spectra, can be also observed in excitation spectra for both molecules. In the high-frequency range above ∼750 cm^–1, the pentacene excitation spectrum differs in appearance from the published emission spectrum,³⁸ in a similar way as previously reported for pentacene on neon clusters.³¹ In this frequency region, there is also no clear correspondence between bands observed in the spectrum of pentacene and the spectra of TIPS-pentacene in Fig. 3. The vibrational modes in this region are typically assigned to collective C–H bending motions. It appears reasonable that these modes are affected substantially by the presence of additional side groups.

We can compare the vibrational frequencies obtained from the TIPS-pentacene excitation spectra with the reported Raman spectra for powder and thin film samples.^13,14,20 We note however, that the Raman spectra measure ground state vibrational frequencies while our spectra reveal vibrational frequencies of the first electronically excited state that are Franck–Condon active. Therefore, the modes observable in our spectra are in principle not limited to Raman active modes. Nevertheless, we find a similar number of vibrational modes than observed in the Raman spectrum. The out-of-plane butterfly modes can, as mentioned above, not be observed in the condensed phase. The in-plane breathing mode appears at similar frequencies in the electronically excited and the ground state, and modes at ∼600 cm^–113,14 and 788 cm^–113,14,20 can also be observed in the Raman spectra. The observed higher vibrational frequencies differ by between 10 and 55 cm^–1 between the two different electronic states and the relatively intense band observed at a vibrational excitation of 1428 cm^–1 in the electronically excited state does not appear in the Raman spectra. It seems plausible that this vibrational mode may simply show significant Frank-Condon overlap for the S₀–S₁ transition, while not being Raman active. Note however, that without an unambiguous assignment of all observed vibrational modes it is not a priori clear if the modes we observe as Franck–Condon active modes in the first electronically excited state do actually correspond to the Raman active modes reported for the ground state.

Vibrations of electronically excited states and the ground state of TIPS-pentacene were previously also reported from 2D electronic spectroscopy measurements^14,20 and transient absorption spectroscopy.¹³ Thereby TIPS-pentacene was used as a thin film or solution. In such experiments, it is not always straight forward to assign observed vibrational coherences to certain electronic states. Previous studies have compared the Fourier spectra of the observed vibrational coherences to ground state Raman spectra and found agreement within their spectral resolution. Musser et al.¹³ could isolate the vibrational coherence signals for an excited state, while averaging their time-domain data far beyond the decay time of the initially excited S₁ state to the triplet manifold. Consequently, they assign this vibrational wave packet to the triplet manifold, populated by an ultrafast singlet fission process that transfers the vibrational coherence from the initially excited singlet state. The vibrational frequencies involved in this triplet state wave packet are also given in Table II. Except for the vibrational frequency above 1500 cm^–1 these frequencies correspond well to those observed in our spectra for the electronically excited singlet state. Similar to the comparison with vibrational modes of the molecular ground state S₀, quantitative agreement between vibrational frequencies in the first excited triplet and singlet states is not expected. Therefore, the surprisingly good agreement between the frequencies we report here, and those observed for the excited (triplet) state coherence of Musser et al.¹³ could indicate strong similarities between the vibrational frequencies of the initially excited singlet and the final triplet state.

It is clear that a short laser pulse with sufficient bandwidth, centred on the S₀–S₁ excitation, as used in previous experiments^13,14,20 will excite a coherent vibrational wave packet containing all the vibrational modes observed in our linear excitation spectrum. Whether all these vibrational excitations are expected to transfer equally well to the triplet manifold in experiments reported by Musser et al.¹³ is unclear. We note, that differences in the coherence lifetimes of various vibrational modes between our isolated molecules and thin film or solution phase samples can affect the number of vibrational modes observed in various experiments. We believe that detailed information on vibrational frequencies coupling to the S₀–S₁ excitation of TIPS-pentacene may aid future assignments of vibrational frequencies observed in coherent electronic spectroscopy techniques and thus contribute to the further study of ultrafast singlet fission processes.

In this study, we have presented a detailed high-resolution LIF spectroscopic study comparing TIPS-pentacene and pentacene embedded in helium nanodroplets and attached to solid rare-gas clusters (neon and argon). The results reveal significant differences in the vibronic structures of these molecules, caused by the additional side groups on the pentacene backbone. While pentacene in helium nanodroplets shows a well-resolved, sharp electronic origin transition, TIPS-pentacene exhibits a broad partially-resolved vibrational progression of very low-frequency vibrational modes (10–20 cm^–1). The corresponding motion seems to be significantly damped by the helium nanodroplet environment, while the modes are not at all observable for TIPS-pentacene on solid rare gas clusters.

We compare the observed vibrational frequencies for the first electronically excited singlet state S₁ to frequencies obtained from Raman spectroscopy for the electronic ground state^13,14,20 and find a similar number of vibrational modes, but as expected no perfect agreement of the vibrational frequencies. We find quantitatively better agreement between the S₁ vibrational frequencies and those previously observed for coherent wave packet in the triplet manifold of TIPS-pentacene thin films.

These findings contribute to a deeper understanding of the vibrational coherences formed in femtosecond experiments of thin film molecular samples. Such coherences have recently been proposed to contribute substantially to ultrafast singlet fission processes, motivating much of the research conducted on TIPS-pentacene and similar organic semiconductor molecules.

We gratefully acknowledge the funding from the Deutsche Forschungsgemeinschaft DFG (Grant Nos. STI 125/25-1 and RTG 2717) and the COST Action CA21101 “Confined Molecular Systems: From a New Generation of Materials to the Stars (COSY)”.

Figure 4 compares the line shapes observed for various vibronic transitions for TIPS-pentacene in helium nanodroplets. The substructure observed in the transition at the lowest transition frequency, attributed to progressions of low-frequency vibrational modes, damped by the helium environment (see main text), are also observed for higher vibrational excitations. Despite lower signal-to-noise ratio the shape is still observable at vibrational excitation frequencies of 257 and 796 cm^–1. For the transition at 1142 cm^–1 the line shape cannot be observed clearly. This may be explained by the absence of the structure for this band or by additional lifetime broadening of these vibronic transitions, or the partial overlap between adjacent bands.