Charge-carrier photogeneration in single-component organic carbazole-based semiconductors via low excitation power triplet-triplet annihilation

It is generally believed that intrinsic charge generation via an autoionization mechanism in pristine single-component organic semiconductors is impossible upon photoexcitation within the lowest excited singlet state due to the large exciton binding energy. However, we present measurements of thermally stimulated luminescence, light-induced ESR, and photocurrent in the carbazole-based molecule 3′,5-di(9H-carbazol-9-yl)-[1,1′-biphenyl]-3-carbonitrile (mCBP-CN) films, revealing that charge-carrier pairs are efficiently produced upon excitation near their absorption edges. The photocurrent measurements show a superlinear dependence on the cw-photoexcitation intensity even at very low excitation power (below $1\mathrm{mW}/{\mathrm{cm}}^2$), suggesting a bimolecular nature of the charge photogeneration process. The photocurrent measured over a broad temperature range of 5–300 K exhibits a prominent maximum at moderately low temperature around 170 K and rolls off significantly at higher temperatures. This correlates remarkably with the maximum of delayed fluorescence induced by bimolecular triplet-triplet annihilation (TTA), i.e., triplet fusion, in this material. This behavior implies that the photocurrent is governed mainly by the TTA-induced production of geminate pairs and only a little by their subsequent dissociation. Moreover, we find that the field-assisted dissociation probability of photogenerated charge pairs becomes almost temperature-independent at temperatures below 100 K. This can be quantitatively described using a charge dissociation model accounting for the energy disorder and the distribution of geminate-pair radii. The key conclusion of this study is that triplet fusion can promote the energy up-conversion (to 5.42 eV), thereby enabling the autoionization of a high-energy neutral excited state. This serves as the predominant mechanism of intrinsic photogeneration in this single-component heavy-atom-free system. We attribute the effect to efficient intersystem crossing in mCBP-CN, a high triplet energy level (2.71 eV), and very long-lived triplet excitations. A broader implication of this finding is that the so far neglected mechanism of TTA-facilitated charge-carrier generation might be relevant for organic long-persistent luminescence materials, and even for organic photovoltaics and potentially for photocatalytic water splitting processes.

Figure 1

(a) Molecular structure of mCBP-CN, (b) schematic energy levels diagram of investigated mCBP-CN single-layer photodiode.

Figure 2

(a) Absorption (orange solid line) where OD denotes optical density, low-temperature steady-state PL (blue solid line), and delayed emission (blue dashed line) of a mCBP-CN film measured at 77 K with a standard spectrofluorometer using a $\mathrm{Xe}$ lamp for excitation. The delayed emission is measured with 50-ms delay time after excitation and an integration time of 100 ms. (b) Absorption (orange solid line) and steady-state PL (blue solid line) measured at 77 K of a rr-P3HT film. (c) Dependence of the integrated TSL intensity upon the photon energy of excitation (red symbols) and PC action spectra (green symbols) of a mCBP-CN film; (d) action spectrum of TSL (red symbols) measured in this work as well as PC (green circles) adopted from Ref. [32] for rr-P3HT film. Dashed lines are a guide to the eye. (e) TSL glow curves of mCBP-CN film and of (f) a rr-P3HT film, both measured after excitation at 5 K with different photon energies of the cw-excitation light (wavelengths of the excitation are indicated in the figure).

Figure 3

Evidence for efficient TTA in mCBP-CN films: (a) dependence of spectrally integrated phosphorescence (Ph) (green triangles) and DF (blue squares) intensity on pulsed nanosecond-laser excitation intensity at 5 K, along with power-law fits to the data (solid line); (b) intensity of spectrally integrated Ph, DF, and PC $(J_{\mathrm{ph}})$ versus temperature (symbols), along with guides to the eye (solid lines); (c) double-logarithmic plot of the PC vs cw-excitation intensity at 350 nm measured at 5 K (red circles). The observed supralinear dependence with slope 1.5 indicated by the red line is principally similar for the diode measured at both short-circuit and reverse bias (−7 V) conditions. (d) Suggested simplified energy diagram for the charge photogeneration process in mCBP-CN films, where IC and ISC denotes internal conversion and intersystem crossing, respectively.

Figure 4

PLDMR triplet spectra of fluorescence (380–420-nm bandpass filter, blue) and fluorescence plus phosphorescence (overall PL spectrum ≥355 nm, red). The bandpass-filtered DF-based spectrum shows increased PLDMR contrast (ΔPL/PL) in comparison to the long-pass-filtered spectrum that also probes phosphorescence. This is evidence that the PLDMR mostly stems from TTA-UC-induced DF rather than phosphorescence. The inset shows the probed Δm_s = ±2 triplet transition as a black arrow, in contrast to the gray arrows for the Δm_s = ±1 transitions.

Figure 5

Light-induced ESR of mCBP-CN under 365-nm excitation at 10 K due to charge carriers [blue lines or symbols in (a),(c),(e)] or triplets [red lines or symbols in (b),(d),(f)]. (a),(b) LESR signal as function of the applied magnetic field. The vertical arrows indicate the signal peaks where transients are measured. The insets show the Zeeman splitting of charge carriers or triplets and the probed Δm_s = ±1 and ±2 ESR transitions, respectively. (c),(d) Illumination-dependent transient signal evolution. The signals grow upon illumination and decay in the dark with drastically different temporal behaviors. (e),(f) Semilog plots of the signal decays in the dark after terminating the excitation. For charges, a logarithmic time dependence $I_{\mathrm{LESR}}\sim -\log (t)$ is observed. Charges accumulate and decay over hours. For the triplet signal, an exponential decay $I_{\mathrm{LESR}}\sim \exp (-(\mathrm{t}/\tau ))$ with a time constant τ = 1.7 s is observed.

Figure 6

(a) Arrhenius plot of the Onsager probability of charge dissociation extracted from the experimental data of Fig. 3 (black circles) and from the analytical hopping model (red curve) with the yield dissociation values normalized to the yield at 290 K. (b) Arrhenius plot of the dissociation probability of a short-distance electron-hole pair (r₀= 1 nm) for a disordered (circles, squares) and a disorder-free (lines) system in the absence of electric field (blue line and squares) and in the presence of an electric field (red line and circles).