Photocatalytic doping of organic semiconductors

Abstract

Chemical doping is an important approach to manipulating charge-carrier concentration and transport in organic semiconductors (OSCs)^1-3 and ultimately enhances device performance^4-7. However, conventional doping strategies often rely on the use of highly reactive (strong) dopants^8-10, which are consumed during the doping process. Achieving efficient doping with weak and/or widely accessible dopants under mild conditions remains a considerable challenge. Here, we report a previously undescribed concept for the photocatalytic doping of OSCs that uses air as a weak oxidant (p-dopant) and operates at room temperature. This is a general approach that can be applied to various OSCs and photocatalysts, yielding electrical conductivities that exceed 3,000 S cm^-1. We also demonstrate the successful photocatalytic reduction (n-doping) and simultaneous p-doping and n-doping of OSCs in which the organic salt used to maintain charge neutrality is the only chemical consumed. Our photocatalytic doping method offers great potential for advancing OSC doping and developing next-generation organic electronic devices.

Fig. 1. Photocatalytic doping concept.

a–d, Schematics of the photocatalytic doping process: weak dopants cannot oxidize or reduce the OSCs (a); PCs in the ground state cannot oxidize or reduce the OSCs (b); PCs in the excited state can oxidize or reduce the OSCs and be regenerated by the weak dopants (c and d). This photocatalytic doping process occurs in thin films and ensures that the molecular packing of the OSCs is preserved after doping. e,f, Proposed photocatalytic oxidation or p-doping (right) and reduction or n-doping (left) cycle (e); chemical structures of the PCs, weak dopants, salts (counterions) and OSCs (f).

Fig. 2. Photocatalytic p-doping of PBTTT.

a, Schematic of the photocatalytic p-doping method: the OSC film is immersed in the PC solution, which also contains TFSI^– counterions, and irradiated with light in the presence of O₂, which acts as the weak p-dopant. The PC solution is recovered and the OSC film is washed with clean solvent and dried in N₂. [C_LiTFSI] and [C_PC] are the concentrations of the LiTFSI and PC solutions, respectively. b, Absorption spectra of undoped and photocatalytically doped PBTTT films. Photocatalytic doping requires both PC and light. c, XPS analysis of undoped and photocatalytically doped PBTTT films reveals a distinct TFSI^– signal in the photocatalytically doped samples. d, S(2p) and F(1s) XPS analysis of undoped and photocatalytically doped PBTTT films. The TFSI^– signal increases with light irradiation time. e, Ultraviolet photoelectron spectroscopy of undoped and photocatalytically doped PBTTT films. The work function of photocatalytically doped PBTTT films increases with light irradiation time. f, Electrical conductivity of undoped and photocatalytically doped PBTTT films demonstrates that photocatalytic doping can happen only when both PC and light are present. Points, mean; error bars, s.d. (not visible); n = 10 independent samples. Source Data

Fig. 3. Mechanism and generality of the photocatalytic p-doping process.

a, Schematic diagram of the transition of the ground state, excited state, reduced state and excited reduced state of Mes-Acr-Me⁺ in the presence of the optically transparent Et₃N in 455 nm irradiation, followed by regeneration in air. b–d, False-colour plots of the in situ transient absorption spectra of Mes-Acr-Me⁺ + 10 equivalents Et₃N, as a function of detector wavelength and delay time: initial state in N₂ atmosphere (b); after 455 nm light irradiation for 120 s in N₂ (c); and regeneration in air (d). Two strong photoinduced absorption peaks (477 nm and 551 nm, b and d) indicate the formation of the excited state of the PC (Mes-Acr-Me^+*). The photoinduced absorption peak at 658 nm (c) indicates the formation of the PC excited reduced state (Mes-Acr-Me^•*). e,f, Electrical conductivity of PBTTT (e) and P(g₄2T-T) (f) photocatalytically doped by different PCs in air. The photocatalytic p-doping conditions are the same as in Fig. 2a. Points, mean; error bars, s.d. (not visible); n = 10 independent samples. g, Electrical conductivity enhancement of OSCs (compared with the undoped films) plotted against the energy barrier of single-electron transfer from OSCs to excited PCs. Source Data

Fig. 4. Photocatalytic n-doping and simultaneous photocatalytic p-doping and n-doping.

a,b, Schematics of the photocatalytic n-doping process: BBL films are covered by the PC solution (0.01 M Mes-Acr-Me⁺ in 3:1 BuOAc:CH₃CN) that also contains 0.1 M [EMIM][TFSI]. The weak n-dopant Et₃N (0.1 M, a) or a physically separate p-type P(g₄2T-T) film immersed in the PC solution (b) is used to regenerate the PC. After irradiation in N₂, the PC solution is recovered, and the OSC films are washed with clean 3:1 BuOAc:CH₃CN solvent and dried in N₂. c, Proposed catalytic cycle for simultaneous photocatalytic p-doping and n-doping of OSCs. The photoactivated PC extracts electrons from the p-type P(g₄2T-T) (or Et₃N) and donates them to the n-type BBL. d, Differential absorption spectra of BBL films after photocatalytic n-doping by Mes-Acr-Me⁺ with Et₃N or P(g₄2T-T). e, Electrical conductivity of undoped and photocatalytically n-doped BBL films. f, Electrical conductivity of simultaneous photocatalytic p-doped P(g₄2T-T) and n-doped BBL. Data in e and f: points, mean; error bars, s.d. (not visible); n = 10 independent samples. Source Data

Extended Data Fig. 1. Photocatalytic doping concept.

a, Photocatalytic oxidation/p-doping cycle. The photocatalyst in the ground state is photoactivated to its excited state and converted to its reduced state upon oxidation of the organic semiconductor. The photocatalyst in the reduced state is activated to its excited reduced state and then regenerated by the weak p-dopant. b, Photocatalytic reduction/n-doping cycle. The photocatalyst is first photoactivated to its excited state and can oxidize the weak n-dopant. Upon oxidation of the latter, the photocatalyst is converted to its reduced state and then photoactivated to obtain an excited reduced state capable of reducing the organic semiconductor.

Extended Data Fig. 2. GIWAXS analysis of PBTTT.

a-d, 2D GIWAXS patterns of undoped PBTTT (a) and doped PBTTT thin films photocatalyzed by different photocatalysts Mes-Acr-Ph⁺ (b), Mes-Acr-Me⁺ (c), and Acr-Me⁺ (d). e-f, Out-of-plane (e) and in-plane (f) GIWAXS line cuts of undoped and doped PBTTT thin films.

Extended Data Fig. 3. Lamellar packing and π-π stacking of PBTTT.

a-d, Lamellar packing diffraction peak analysis of undoped PBTTT (a) and doped PBTTT thin films photocatalyzed by Mes-Acr-Ph⁺ (b), Mes-Acr-Me⁺ (c), and Acr-Me⁺ (d). e-h, π-π stacking diffraction peak analysis of undoped PBTTT (e) and doped PBTTT thin films photocatalyzed by Mes-Acr-Ph⁺ (f), Mes-Acr-Me⁺ (g), and Acr-Me⁺ (h). i-k, Packing distances (i), coherence length (j), and paracrystalline disorder (k) of lamellar packing and π-π stacking of PBTTT. Photocatalyzed doped PBTTT shows longer lamellar packing distance, shorter π-π stacking distance, and more regular π-π stacking than undoped PBTTT.

Extended Data Fig. 4. Effect of thickness and light irradiation power.

a, Electrical conductivity of doped PBTTT (photocatalyzed by Acr-Me⁺) films with various thicknesses. Points, mean; error bars, s.d. (not visible); n = 10 independent samples. b, Absorption spectra of undoped and doped PBTTT films with various thicknesses. The doped PBTTT shows similar maximum conductivity and polaron absorption intensity regardless of the film thicknesses, suggesting that photocatalytic doping is a bulk doping process. c-d, Electrical conductivity of doped PBTTT (photocatalyzed by Acr-Me⁺) films under different light exposure power, plotted as a function of light irradiation time (c) and light irradiation dose (d). Points, mean; error bars, s.d. (not visible); n = 10 independent samples.

Extended Data Fig. 5. TAS of Mes-Acr-Me⁺.

a, Schematic diagram of the transitions from the ground state to the excited state and the intramolecular charge-transfer (ICT) excited state of Mes-Acr-Me⁺, under 390 nm laser pump pulse. b, False color representation of the transient absorption spectra of Mes-Acr-Me⁺ solution (0.01 M, with 0.1 M of LiTFSI, in BuOAc:CH₃CN = 3:1) as a function of delay time and detector wavelength. c, Transient absorption spectra of Mes-Acr-Me⁺ showing the photobleaching (peak center 601 nm) and two significant photoinduced absorption peaks (I, 484 nm; II, 549 nm). d, Time-dependent photobleaching and photoinduced absorption of Mes-Acr-Me⁺, showing immediate raising (< 1 ps, limited by instrumental time resolution) for photobleaching, and immediate raising, slow falling, and slower again raising (raising time ~ 26 ps), indicating the formation of the ICT excited state. e, Time-dependent photoinduced absorption of Mes-Acr-Me⁺, showing a long excited lifetime of 10 ns. f, Time-dependent photobleaching of Mes-Acr-Me⁺, showing a photobleaching time constant of 1.8 ns.

Extended Data Fig. 6. In-situ TAS of Mes-Acr-Me⁺/Et₃N.

a, Schematic diagram of the transition from the ground state to the excited state to the reduced state and excited reduced state of Mes-Acr-Me⁺. b-h, False color representation of the in-situ transient absorption spectra of Mes-Acr-Me⁺ solution (0.01 M, with 0.1 M of Et₃N, 0.1 M of LiTFSI, in BuOAc:CH₃CN = 3:1) as a function of delay time and detector wavelength, with different 455 nm light irradiation time: initial state 0 s (b), 15 s (c), 60 s (d), 90 s (e), 120 s (f), 180 s (g), and regenerated in air (h). The transient absorption spectra reveal two strong photoinduced absorption peaks (477 nm and 551 nm) in the initial and regenerated Mes-Acr-Me⁺, showing the ICT excited state (Mes-Acr-Me^+*). The new photoinduced absorption peak at 658 nm is ascribed to the excited reduced state (Mes-Acr-Me^•*). The transient absorption spectra recovered almost completely to the initial state after exposure to air, suggesting the regeneration of Mes-Acr-Me⁺ by air oxidation. i, In-situ transient absorption spectra of the Mes-Acr-Me⁺/Et₃N solution in the initial state, after 455 nm light irradiation time of 120 s, and regenerated in air. The transient absorption spectra present the formation of a reduced state (Mes-Acr-Me^•) and excited reduced state (Mes-Acr-Me^•*), and regeneration to Mes-Acr-Me⁺ by air oxidation. j-k, Time-dependent photoinduced absorption, showing similar features in the initial and regenerated states of Mes-Acr-Me⁺/Et₃N solution, indicating the same ICT excited state formation (Mes-Acr-Me^+*). l-n, Time-dependent photoinduced absorption of Mes-Acr-Me⁺/Et₃N solution after irradiation at 455 nm for 0 s (l) and 120 s (m), and regeneration in air (n), showing that the excited lifetime of Mes-Acr-Me⁺ is ~1 ns (in the presence of 10 eq Et₃N), and excited lifetime of Mes-Acr-Me^• is 70 ps.

Extended Data Fig. 7. Photocatalytic p-doping mechanisms.

a, Photocatalytic p-doping cycle of PBTTT in the presence of Acr-Me⁺ and O₂. b, Corresponding DFT-calculated Gibbs free energy profiles.

Extended Data Fig. 8. Generality of photocatalytic doping: absorption spectra.

a-f, Absorption spectra of undoped and doped (photocatalyzed by Acr-Me⁺/air) P(g₄2T-T) (a), P(g₄2T-TT) (b), P3HT (c), PBTTT (d), PTQ1 (e), and gDPP-g2T (f) films.

Extended Data Fig. 9. Generality of photocatalytic doping: electrical conductivity.

a-d, Electrical conductivity of PTQ1 (a), P3HT (b), P(g₄2T-TT) (c) and gDPP-g2T (d), the thin films were photocatalytically doped by different photocatalysts. Points, mean; error bars, s.d. (not visible); n = 10 independent samples.

Extended Data Fig. 10. Use of simultaneous photocatalytic p/n-doping to build thermoelectric generators.

a, Current of a planar thermoelectric module integrating one P(g₄2T-T) p-leg (W = 40 mm, L = 2 mm, d = 20 μm) and one BBL n-leg (W = 20 mm, L = 2 mm, d = 10 μm) with silver contacts as a function of time for different load resistances (1-10⁷ Ω) and temperature gradients (∆T = 10-50 K). The devices show a stable, fast, and fully reversible response. The thermoelectric module was fabricated by drop-casting the p-type polymer P(g₄2T-T) and n-type polymer BBL on a 25-μm polyethylene terephthalate (PET) substrate using a shadow mask. The P(g₄2T-T) and BBL legs were doped by simultaneous photocatalytic p/n-doping using Mes-Acr-Me⁺ as the PC. A silver paste was used to print the contacts. The devices were tested in nitrogen. b, Voltage output of the thermoelectric module as a function of the load current (current output). c, Power output as a function of the resistance load recorded at different ∆T. d, Power output of the planar thermoelectric module as a function of ∆T.