Electron-Ion Coupling Breaks Energy Symmetry in Bistable Organic Electrochemical Transistors

Abstract

Organic electrochemical transistors are extensively studied for applications ranging from bioelectronics to analog and neuromorphic computing. Despite significant advances, the fundamental interactions between the polymer semiconductor channel and the electrolyte, which critically determine the device performance, remain underexplored. Here, we examine the coupling between the benchmark semiconductor PEDOT:PSS and an ionic liquid to explain the bistable and non-volatile behavior observed in OECTs. Using X-ray scattering and spectroscopy techniques, we demonstrate how the electrolyte modifies the channel composition, enhances molecular order, and reshapes the energetic landscape. Notably, the observed bistability arises from asymmetric and path-dependent energetics during doping and dedoping, resulting in two distinct paths, driven by a direct interaction between the electronic and ionic charge carriers. These findings highlight the electrolyte's role in tuning charge carrier dynamics, positioning it as a powerful yet underutilized lever for enabling novel device functionalities.

Figures (4)

• Figure 1: OECTs with PEDOT:PSS channel and [EMIM][EtSO4] electrolyte. (a) Generalized layout of an OECT with PEDOT:PSS channel. Devices in this work typically feature a capacitive side-gate (Fig. \ref{['fig:S_Devices']}). (b) PEDOT:PSS is a heterogeneous material, where the electronic charge transport through aggregated, paracrystalline entities is bridged by amorphous regions. (c) Transfer curve in linear and logarithmic scales of a bistable OECT based on PEDOT:PSS with an [EMIM][EtSO4] solid-state electrolyte, as reported in ref. bongartz2024bistable ($V_\mathrm{DS}=-200\,mV$). (d--f) Transfer curves of OECTs with PEDOT:PSS channel and plain [EMIM][EtSO4], [EMIM][TFSI], and $100\,m\molar$ aqueous NaCl electrolyte ($V_\mathrm{DS}=-10\,mV$). (g) Operating an OECT ([EMIM][EtSO4] electrolyte) with a gate current instead of a gate voltage ($V_\mathrm{DS}=-10\,mV$). (h) The folded $I_\mathrm{D}$-response shows no hysteresis. Instead, the hysteresis reflects in the potential difference between source and gate.
• Figure 2: Semiconductor analysis. (a--c) XPS spectra (raw and fitted) of pristine and electrolyte-treated PEDOT:PSS thin films with core levels (a) Na 1s, (b) N 1s, and (c) O 1s. See Fig. \ref{['fig:S_XPS_C_S']} for C 1s and S 2p spectra. Raw data were fitted to Gauss--Lorentz profiles. (d) Raman spectra (high-energy range) measured with $\lambda_\mathrm{exc}=532\,nm$. The full spectra and peak assignment are provided in Fig. \ref{['fig:S_Raman']} and Tab. \ref{['tab:S_Raman']}. (e) DFT calculations of distorted and planar PEDOT reproduce the experimental observations in (d). (f) Schematic illustrating the effect of [EMIM][EtSO4] on PEDOT:PSS. Along with an ion exchange, where PSS is removed and [EMIM]+ and [EtSO4]- are retained, the PEDOT chains reorganize and self-assemble with higher order.
• Figure 3: In-operando doping studies. (a, b) Absorbance spectra (discharging cycle) of PEDOT:PSS with [EMIM][EtSO4] and [EMIM][TFSI] electrolyte. [EMIM][EtSO4] produces feature-rich spectra, reflecting high aggregation, while [EMIM][TFSI] suggests an amorphous nature. (c) Absorbance spectra are fitted using a vibronic model that incorporates the lowest five vibronic transitions (0--0 through 0--4). This approach allows to infer the absorbance share owing to undoped PEDOT in aggregates and amorphous regions. (d) From this, the aggregate fraction ($AF$) follows for each potential. The omitted potential regime cannot be reliably fitted due to dominant polaron absorbance (\ref{['Note_S:Spectroelectrochemistry_Fits']}). Insets: Differential absorbance for the dis-/charging step between potentials of $-0.2\,V$ and $-0.15\,V$, revealing a stark asymmetry in the spectral signatures.
• Figure 4: Energetics of the doping cycle. (a) Charge carrier density ($n$) as a function of applied potential, calculated by integrating the current with respect to time (Fig. \ref{['fig:S_CA']}). See also Fig. \ref{['fig:S_CA_Diff_EMIMEtSO4']}. (b) $E_{00}$ energies extracted from absorbance fits (Fig. \ref{['fig:3']}c), showing a marked asymmetry between doping and dedoping. (c, d) Aggregate fraction ($AF$) and $E_{00}$ energies for samples with [EMIM][TFSI] and NaCl electrolyte. (e) Schematic illustration of the ionic and electronic population during doping. Left: Fully dedoped state; abundant [EMIM]+ and only residual [EtSO4]- in ordered domains. Center: low-doping regime; initial hole injection into highly ordered PEDOT aggregates, partially co-stabilized by [EtSO4]-. This stabilization must be overcome during dedoping. Right: high-doping regime; after aggregate saturation, holes extend into the amorphous phase.