Traps and transport resistance are the next frontiers for stable non-fullerene acceptor solar cells

The study addresses the critical challenge of operational stability in high-performance organic solar cells (OSCs), particularly those employing non-fullerene acceptors (NFAs) such as the PM6:Y6 system. While power conversion efficiencies have surpassed 18%, device stability under realistic conditions remains limiting. Degradation in OSCs can arise from environmental exposure (oxygen, water), illumination, thermal stress, materials morphology, and interfacial processes. The authors aim to elucidate the dominant degradation pathways in inverted PM6:Y6 devices and identify the mechanisms underlying thermally induced losses. They observe two distinct pathways: (1) a short-circuit current (Jsc) loss requiring both light and oxygen, and (2) a thermally driven degradation characterized by reductions in open-circuit voltage (Voc) and fill factor (FF), independent of oxygen or illumination. The central hypothesis tested is that thermal aging primarily induces trap state formation in the active layer, which increases non-radiative recombination and transport resistance, thereby limiting Voc and FF. Understanding and mitigating trap formation is proposed as a key frontier for achieving long-lived, high-performance OSCs.

Prior work has shown that oxygen and moisture can react under illumination to degrade OSC components, including oxidation of interlayers/electrodes and generation of extraction barriers. Photoinduced radicalization of molecular oxygen can cause photobleaching and increased energetic disorder. In fullerene-based devices, photoinduced fullerene dimerization under inert conditions leads to Jsc losses. NFAs were pursued to avoid fullerene dimerization, yet NFAs can also dimerize depending on donor–acceptor energetics via back-transfer or intersystem crossing. Dominant NFA degradation mechanisms include intermixing/demixing of blend components and structural decomposition (e.g., IT-4F) near ZnO and oxygen, with molecular planarity sometimes improving stability. For PM6:IT-4F, similar degradation mechanisms occur under thermal dark aging and room-temperature illumination (without UV). Reports also indicate photocatalytic decomposition at ZnO interfaces in ITIC-based devices and light-induced ZnO-related degradation. This body of literature underscores the roles of oxygen/light-induced chemical processes and morphology in stability, motivating deeper investigation into NFA systems like PM6:Y6 under thermal stress.

Device fabrication: Inverted architecture glass/ITO/ZnO/PM6:Y6/MoO3/Ag. Active layer PM6:Y6 (1:1.2) in chloroform with 0.5 vol.% 1-chloronaphthalene; spin-coated to 100 or 200 nm, annealed at 100 °C (10 min). ZnO nanoparticle layer spin-coated and annealed at 200 °C (30 min). 10 nm MoO3 and 100 nm Ag thermally evaporated. Device area ~10.4 cm². Aging protocols (all at open circuit): combinations of air vs nitrogen, room temperature vs 85 °C, dark vs illumination (warm-white LEDs without UV, ~0.5 sun-equivalent). Main focus: thermal aging at 85 °C in N2, dark. Devices characterized fresh, after 24 h, and after 96 h. Electrical/optical characterization: Illuminated and dark J–V (AM1.5), suns–Voc and pseudo-FF, capacitance–voltage (50 kHz, 20 mV), impedance spectroscopy and intensity-modulated spectroscopy (IMPS/IMVS) under varied light intensities, temperature-dependent Voc, sensitive EQE (photocurrent spectroscopy), electroluminescence (EL) spectra (constant current drive), photothermal deflection spectroscopy (PDS) for sub-gap absorption, variable angle spectroscopic ellipsometry, and GIWAXS (beamline ALS 7.3.3) for morphology on Si and ZnO/ITO substrates. UPS on devices (after Ag removal) to assess work function and energetic alignment. Modeling: Drift–diffusion simulations (gpvdm) with time-domain Shockley–Read–Hall trapping formalism extended to frequency domain. Global fitting to dark and illuminated J–V, CV, suns–Voc (including Jsc vs light intensity), and impedance spectra to extract trap density, capture cross-sections, and effective carrier mobility as functions of aging. Transport resistance quantified by comparing light J–V to shifted suns–Voc curves; effective single-carrier conductivity near Voc computed as σ = 2ΔV(J)/(J L). Intrinsic carrier density estimated from chemical capacitance via n(Vac)=n(V)+∫ C(V)dV and ni = n(V) exp(−V/2kT).

• Two distinct degradation pathways identified in PM6:Y6 inverted cells:
  1. Illumination-in-air pathway: significant Jsc loss (~30% drop after 96 h) with stable Voc and FF; requires both light and oxygen, consistent with triplet exciton recombination forming reactive oxygen species leading to photobleaching.
  2. Thermal pathway: at 85 °C (air or N2, dark or light), strong Voc and FF losses with Jsc largely unchanged. Voc decreased from ~825 mV (fresh) to ~780 mV (24 h) and <760 mV (96 h); FF declined from ~62% to <50% (24 h) and <40% (96 h). Pseudo-FF remained ~82% across conditions, indicating transport, not recombination competition, limits FF.
• Charge generation unaffected by thermal aging: Jsc stable; transient absorption shows unchanged Y6 exciton dynamics and charge generation; PDS/EQE/EL sub-gap spectra nearly identical before/after aging; Urbach energy ~21.5 meV (200 nm) and ~24.5 meV (100 nm) unchanged.
• Interfacial energetics stable: UPS shows no significant change (~±0.1 eV) in work function or ionization potential; built-in potential unchanged. Suns–Voc ideality factors indicate Voc limited by surface recombination at 1 sun for both fresh and aged.
• Non-radiative recombination increases with aging: Relative EL indicates ΔVoc,nonrad increases by ~80 mV (24 h) and ~100 mV (96 h), matching Voc drop; dark saturation current J0 increases accordingly.
• Effective bandgap considerations: Apparent effective bandgap from Voc(T→0 K) ~1.1 eV; correcting for S1 repopulation barrier (ΔE ≈ 0.12 eV) yields Egeff ≈ 1.22 eV; neither value changes with aging, suggesting radiative losses unchanged.
• FF limited by transport resistance: Voltage loss ΔV between light J–V and shifted suns–Voc curves increases with aging; effective conductivity near Voc for 200 nm devices drops from 4.0×10⁻4 S m⁻1 (fresh) to 2.2×10⁻4 S m⁻1 (24 h) and 1.7×10⁻4 S m⁻1 (96 h). Similar trends for 100 nm devices consistent with weaker FF losses at smaller thickness.
• Trap state formation evidenced by experiment and modeling: Simulations indicate bulk trap density rises from ~10²⁴ m⁻³ (fresh) to ~10²⁶ m⁻³ (96 h), with modified capture cross-sections; effective carrier mobility near Voc decreases from ~5×10⁻8 to ~1–2×10⁻8 m² V⁻1 s⁻1 (96 h). CV under illumination shows increasing capacitance at forward bias with aging; intrinsic carrier concentration ni increases from ~1×10¹5 to ~5×10¹5 m⁻3 (200 nm) over 96 h, consistent with deeper/more numerous traps.
• Morphology: GIWAXS on ZnO/ITO substrates shows subtle aging effects for Y6 but PM6 exhibits increased lamellar stacking fraction with a slight reduction of paracrystalline disorder in lamellar stacking; however, π–π stacking paracrystallinity disorder parameter g increases (from 13.5±0.3% to 16.3±0.5%), approaching neat PM6 (17.4±0.3%), implying reduced π–π stacking quality that can degrade transport. On Si substrates, Y6 ordering increases strongly with aging, but ZnO suppresses strong morphological changes. Overall, thermal aging induces bulk traps that elevate non-radiative recombination (reducing Voc) and increase transport resistance (reducing FF), while charge generation and interfacial energetics remain largely stable.

The separation of degradation pathways clarifies that light-and-oxygen-induced processes primarily diminish Jsc via photochemical reactions, whereas thermal stress induces bulk trap formation that impacts Voc and FF. The constancy of pFF and suns–Voc-derived recombination characteristics under 1 sun, along with stable sub-gap absorption and interfacial energetics, indicates that recombination competition and band-edge disorder do not drive the observed FF loss. Instead, increased trap density lowers effective mobility and conductivity, decoupling applied from internal voltage and elevating transport resistance. The increased ΔVoc,nonrad inferred from EL corroborates enhanced trap-assisted recombination limiting Voc. Morphological analysis suggests that, in the PM6 phase, reduced π–π stacking quality accompanies a greater fraction of lamellar stacking, likely contributing to energetic/structural disorder that promotes trapping and impedes transport. The integrated drift–diffusion modeling reproduces the experimental J–V, CV, and impedance data by increasing trap density and adjusting capture cross-sections while keeping free-carrier mobility parameters otherwise consistent, reinforcing the central role of traps. These insights highlight trap suppression and transport optimization as primary levers for improving thermal stability in PM6:Y6 and likely other NFA-based OSCs.

This work identifies and disentangles two dominant degradation pathways in inverted PM6:Y6 organic solar cells. Under illumination in air, Jsc losses arise from oxygen- and light-driven photochemical processes, whereas thermal aging at 85 °C primarily degrades Voc and FF through aging-induced trap state formation in the bulk active layer. Interfacial energetics, optical bandgap, and charge generation remain largely unaffected. The increasing trap density elevates non-radiative recombination (reducing Voc) and increases transport resistance via reduced effective mobility and conductivity (reducing FF), making transport resistance the major limiter of FF. Controlling trap formation and distribution—by tuning chemical structure, morphology (especially π–π stacking quality in the polymer phase), and processing—emerges as a crucial frontier for enhancing OSC stability and device lifetimes. Future research should focus on correlating molecular design and blend morphology with trap spectra/densities, developing strategies to suppress trap formation during thermal stress, and engineering transport to mitigate trap-induced resistance while maintaining high photogeneration and favorable interfacial energetics.

• The detailed chemical/structural origin of thermally induced trap formation is inferred (e.g., changes in PM6 π–π stacking quality) but not fully resolved; sub-gap EQE dominated by Y6 may mask contributions from PM6 or the PM6:Y6 interface.
• Aging studies were limited to up to 96 h and specific conditions (notably LEDs without UV at ~0.5 sun for illuminated aging); long-term and UV-including conditions were not assessed here.
• Results pertain to the PM6:Y6 system in an inverted architecture with ZnO and MoO3/Ag contacts; generalization to other materials/architectures requires further validation.
• GIWAXS indicates substrate-dependent morphology evolution (Si versus ZnO/ITO), which may complicate direct correlation to in-device morphology and transport.