Oxygen-catalysed sequential singlet fission

Abstract

Singlet fission is the photoinduced conversion of a singlet exciton into two triplet states of half-energy. This multiplication mechanism has been successfully applied to improve the efficiency of single-junction solar cells in the visible spectral range. Here we show that singlet fission may also occur via a sequential mechanism, where the two triplet states are generated consecutively by exploiting oxygen as a catalyst. This sequential formation of carriers is demonstrated for two acene-like molecules in solution. First, energy transfer from the excited acene to triplet oxygen yields one triplet acene and singlet oxygen. In the second stage, singlet oxygen combines with a ground-state acene to complete singlet fission. This yields a second triplet molecule. The sequential mechanism accounts for approximately 40% of the triplet quantum yield in the studied molecules; this process occurs in dilute solutions and under atmospheric conditions, where the single-step SF mechanism is inactive.

Fig. 1

Overview of reaction mechanisms and molecular structures. a Reaction scheme for homogeneous singlet fission. Upon excitation, a singlet exciton interacts with a ground-state chromophore to yield two triplet states. b Photocycle for oxygen-catalysed sequential singlet fission. The excited singlet chromophore transfers part of its energy to ³O₂ in a spin-allowed process, which yields the reactive ¹O₂ species and the triplet state of the SF chromophore. Next, an additional ground-state chromophore is sensitised by ¹O₂ in a heterogeneous singlet fission process, regenerating ³O₂. On the whole, sequential SF forms two triplet states from one absorbed photon, just as in homogeneous SF. c Chemical structures of TIPS-pentacene and TDCl₄. Their first absorption band is located between 600 and 700 nm, and their most stable triplet state is located at approximately half the S₀S₁ bandgap, which is nearly isoenergetic with ¹O₂

Fig. 2

Reaction mechanisms and expected concentration dependence. a Scheme for the mechanism of oxygen-catalysed singlet fission of TIPS-Pn in solution with rate constants. An identical mechanism is proposed for TDCl₄. Owing to the relative energies of S₁ and T₁ compared to ¹O₂, the individual steps of sequential singlet fission are exothermic. b Scheme for the mechanism of homogeneous singlet fission in TIPS-Pn with corresponding rate constants. c Upper panel: expected chromophore concentration dependence for the quantum yields of homogeneous (red) and sequential (green) SF. Lower panel: expected chromophore concentration dependence for rate constants of the individual processes. The association of an excited and ground-state chromophores (k_SF) is the rate-limiting step of homogeneous SF. SF is outweighed by concentration independent, non-SF relaxation pathways (i.e., fluorescence, internal conversion and, to a minor extent, intersystem crossing) for low chromophore concentrations and in the absence of oxygen (black, ${\mathit{k}}_{\mathbf{tot}}\approx {\mathit{k}}_{\mathbf{R}}$). At higher chromophore concentrations, ${\mathit{k}}_{\mathbf{tot}}\approx {\mathit{k}}_{\mathbf{SF}}\left[{\mathit{S}}_0\right]$ (black). In contrast, the rate of the first step in sequential SF, that is, the energy transfer (${\mathit{k}}_1\left[{}_{}^3{\mathbf{O}}_2\right]$), is not affected by the chromophore concentration because ³O₂ is in excess compared to the excited chromophore. However, the subsequent heterogeneous singlet fission (blue, ${\mathit{k}}_2\left[{\mathit{S}}_0\right]$) exhibits a pseudo-first-order reaction rate and thus a linear dependence on the chromophore ground-state concentration. Finally, regardless of how the triplet is formed, its relaxation rate remains unaffected by the chromophore concentration (yellow, k_T). Source Data are provided as a Source Data file

Fig. 3

Spectral evolution in time and oxygen dependence of kinetic traces. a Spectral evolution of a 0.5 mM solution of TIPS-Pn in THF. In the initial 6 ns, the transient spectrum is determined by S₁ excited-state absorption with a maximum at 450 nm. Additionally, signatures of ground-state bleach and stimulated emission are observed at 640 and 700 nm, respectively, in good agreement with the absorption and emission spectra (black and blue lines). After 50 ns, a triplet ESA with a maximum at 500 nm can be observed in lieu of the singlet. No further spectral changes are observed. b Selected transients at 500 nm probe wavelength. The kinetic traces for a 0.5 mM solution of TIPS-Pn (see inset) show a stark contrast when using THF stored under atmospheric conditions (blue) compared to degassed THF (red). At the maximum of the triplet ESA (500 nm), the initial fast decay of the singlet is observed. In the case of degassed THF, the amplitude of the triplet is very weak at µs delays. However, in the solution containing oxygen, after the fast singlet decay, the triplet amplitude rises slowly until a local maximum is observed at a probe delay of ≈ 0.4 µs. This behaviour indicates an oxygen-catalysed SF process. Additionally, the triplet decays much faster than in the degassed solution (${\mathit{\tau }}_{\mathbf{deg}}=28.2\mathbf{\mu s}$; ${\mathit{\tau }}_{{\mathit{O}}_2}=1.8\mathbf{\mu s}$) due to further interaction with oxygen (see Supplementary Note 5 for more details). Source Data are provided as a Source Data file

Fig. 4

Kinetic traces of triplet ESA at selected concentrations with respective fit traces. The kinetic traces at a probing wavelength of 500 nm, that is, the maximum of the triplet ESA, are sensitive to the chromophore concentration. No µs local maximum is visible for the lowest concentration (0.02 mM), as sequential SF and triplet decay occur at comparable rates. At higher concentrations, the S₀ sensitisation process is accelerated, which explains a local maximum on the µs timescale and justifies the shift of that maximum towards shorter probe delays as the chromophore concentration is increased. For concentrations ≥ 10 mM, no maximum is visible as (i) the ¹O₂ sensitisation is the rate-limiting step and (ii) the rate of homogeneous SF surpasses the rate of sequential SF. Source Data are provided as a Source Data file

Fig. 5

Singlet fission (SF) quantum yields and rate constants obtained by a global fit. The upper panel shows the acene concentration dependence of the SF quantum yield. The lower panel analyses the concentration dependence of the three rate constants explaining the time evolution of the singlet and triplet concentration. The experimental results perfectly reproduce the trend expected for the model of catalysed SF. For c ≤ 1 mM, k_R has a concentration-independent value of (8.57 ± 0.19) × 10⁷/s, whereas for c ≥ 10 mM, a linear fit of k_tot (black) yields a slope of (8.98 ± 0.36) × 10⁹/Ms, which is attributed to k_SF. As expected from the model (Fig. 4), an intermediate component— attributed to heterogeneous singlet fission—can be observed only in a concentration range of 0.1 mM < c < 10 mM. The slope of the linear fit provides a value of k₂ = (1.62 ± 0.19) × 10¹⁰/Ms. This value is in very good agreement with reported oxygen quenching rates. The triplet decays with a concentration-independent rate of k_T = (6.55 ± 0.30) × 10⁵/s. Source Data are provided as a Source Data file

Fig. 6

Relative energy differences for selected SF molecules. N₂-Pn, N₄-Pn, TDT, TDF₄, TDCl₄ and TIPS-Pn stand for diaza-TIPS-pentacene^,, tetraaza-TIPS-pentacene^,, phenazinothiadiazole, tetrafluorphenazinothiadiazole, tetrachlorphenazinothiadiazole and TIPS-tetracene, respectively. The S₁ – T₁ (red) and T₁– S₀ (green) energy gaps are shown and compared with the ¹O₂ – ³O₂ lowest energy gap of molecular oxygen (black). The feasibility of the two steps of catalysed SF, namely, ³O₂ and S₀ sensitisation, depends on the relative energies of the states involved. In the first step, the excited S₁ chromophore transfers part of its excitation energy to ³O₂. This reaction can occur only if E(S₁– T₁) ≥ E(¹O₂– ³O₂), i.e., if this energy difference is above the black dashed line. In the second step, a ground-state chromophore can be excited by energy transfer from ¹O₂ if E(T₁ – S₀) ≤ E(¹O₂ –³O₂), i.e., if this second energy gap is below the dashed black line. These prerequisites are fulfilled for many acenes exhibiting SF. According to this model, heterogeneous SF should not occur in TIPS-Tn, as the T₁ state lays higher than ¹O₂ in energy (see Supplementary Fig. 7 for measurements done with and without O₂). Pentacene would in principle fulfil all conditions, but it is oxidised to an endoperoxide and thus not a suitable candidate