Elimination of charge-carrier trapping by molecular design

Organic semiconductors often exhibit poorer charge transport than inorganic counterparts due to two fundamental limitations: (1) low carrier mobility stemming from weak van der Waals and π–π interactions that lead to energetic and structural disorder and hopping transport; and (2) trapping of charge carriers by extrinsic impurities so that only a fraction of injected carriers contributes to conduction. While mobilities >10 cm^2 V^−1 s^−1 have been achieved by optimizing packing, trapping of either electrons or holes remains the main cause of imbalanced transport. A universal trap-free window has been identified: ambipolar trap-free transport occurs when EA ≥ 3.5 eV and IE ≤ 6.0 eV, implicating common extrinsic traps (oxygen for electrons, water clusters for holes). A key open question is whether intrinsic trap-free transport for both carriers is achievable in organic semiconductors whose band gap exceeds this 2.5 eV window (as in blue OLED materials), where either HOMO or LUMO (or both) necessarily lies outside the window. This limitation has hindered efficient single-layer blue OLEDs and leads to imbalanced transport and reduced device lifetime in multilayer architectures. The authors propose and demonstrate a molecular design strategy using donor–acceptor molecules with spatially separated HOMO (on donor) and LUMO (on acceptor), where stacking is tuned so donor groups shield the acceptor core, blocking impurity interactions with the LUMO to suppress electron trapping. This bottom-up concept aims to eliminate the detrimental effects of external impurities in wide-band-gap organic semiconductors.

Prior work established that charge transport in organic semiconductors is governed by hopping in disordered systems (Bässler, Brédas et al.) and that extrinsic impurities (oxygen and water) create traps that limit electron and hole transport (Nicolai et al.; Haneef et al.; Zuo et al.). A universal energetic window (EA ≥ 3.5 eV; IE ≤ 6.0 eV) was identified enabling trap-free transport across polymers and small molecules (Kotadiya et al.). High-mobility, trap-free transport has been observed in specific conjugated polymer systems and by dilution strategies (Nikolka et al.; Abbaszadeh et al.). Triazine-based materials are known efficient electron-transport hosts in OLEDs (Chen et al.). Carbazole–triazine donor–acceptor combinations have been used as TADF blue emitters, with more donor units improving OLED efficiency, though their individual charge-transport properties were not addressed (Lee et al.). The present work builds on these insights by targeting molecular packing and spatial shielding to mitigate extrinsic trapping beyond what is predicted by energetics alone.

Materials and molecular series: Two series of blue-emitting donor–acceptor molecules were synthesized. Series 1 varies the number of carbazole (Cz) donors linked via phenylene to a triazine (Trz) acceptor: 1CzTrz, 2CzTrz, 3CzTrz, 4CzTrz, 5CzTrz. Fluorinated analogues 1CzTrz-F to 3CzTrz-F (with F on carbazole) were also made. Series 2 varies fluorination on carbazole in DTPT-DCz, DTPT-DFCz, DTPT-D2FCz with a di-tert-butyl triazine core and two carbazoles. All synthesized per literature procedures and purified by vacuum sublimation.

Energy level characterization: Ionization energies (IE) measured by UPS; electron affinities (EA) derived from cyclic voltammetry (CV), and optical gaps from UV–vis/PL spectroscopy. The 1–5CzTrz series showed IE ≈ −5.8 eV and EA ≈ 3.1 ± 0.1 eV; fluorinated 1–3CzTrz-F had EA 2.7–2.9 eV. DTPT series showed EA increase from 2.6 eV (DTPT-DCz) to 2.8 eV (DTPT-D2FCz) with increasing fluorination.

Device fabrication and electrical measurements: Electron-only diodes were fabricated on glass. Substrates were cleaned (detergent, ultrasonication in acetone and IPA), baked at 140 °C (10 min), UV-ozone treated (20 min), and transferred to N2 glovebox. Stack: 30 nm Al bottom electrode, organic layer ~80–118 nm (depending on compound), 4 nm TPBi tunnel barrier, 5 nm Ba/100 nm Al top electrode. Ohmic electron injection achieved via TPBi tunnel barrier decoupling. J–V characteristics were measured under N2 using Keithley 2400. Thickness and temperature dependencies were checked to rule out injection barriers or V_bi effects.

Modeling of transport and trapping: J–V curves were modeled using a drift–diffusion framework. For compounds exhibiting quadratic J ∝ V^2 (trap-free SCLC), electron mobility was extracted (3CzTrz: μ_e ≈ 2×10^−3 m^2 V^−1 s^−1). Other compounds were modeled by adding electron traps with Gaussian energy distributions; trap densities and parameters provided in supplementary tables.

Computational DOS and trap energetics: Density of states (DOS) for EAs of amorphous and crystalline phases of the molecules and molecular oxygen were computed (details in Supplementary Information). Calculations compared EAs of organic LUMOs and O2 to assess trap depth; electrostatic and polarization contributions analyzed. Electronic transfer integrals between close O2–host pairs were computed to evaluate coupling strength and trap effectiveness.

Structural characterization and morphology: Single-crystal X-ray diffraction (XRD) was performed on 1CzTrz-F, 3CzTrz-F, DTPT-DCz, and DTPT-D2FCz to determine packing motifs, dimer formation, torsion angles, plane-to-plane distances, and donor–acceptor spatial arrangements. Magic-angle spinning (MAS) solid-state NMR (1H, 19F) on vapor-deposited thin films probed local ordering and symmetry breaking around fluorinated sites, indicating non-amorphous local packing. NMR details: 1H at 850.27 MHz, 50 kHz MAS; 19F at 470.61 MHz, 25 kHz MAS; direct excitation with specified pulse lengths and recycle delays.

Additional measurements: Photoelectron spectroscopy (AC-2), MALDI-TOF/ToF and APCI MS, and CV (three-electrode cell in acetonitrile with 0.05 M n-Bu4NPF6, 100 mV s^−1, under Ar). Data availability and computational force fields are referenced via Figshare DOIs.

• Despite similar donor/acceptor chemistry and EAs outside the trap-free window, electron transport varies by 4–5 orders of magnitude across 1–5CzTrz and 1–3CzTrz-F. An optimum occurs at three carbazole donors (3CzTrz and 3CzTrz-F), exhibiting nearly trap-free, quadratic SCLC (J ∝ V^2).
• For 3CzTrz, trap-free electron mobility extracted from quadratic regime is ~2×10^−3 m^2 V^−1 s^−1. In separate balanced electron- and hole-only measurements on 3CzTrz, both carrier mobilities are nearly trap-free and balanced at ~2×10^−4 m^2 V^−1 s^−1.
• 3CzTrz outperforms TPBi in electron transport: TPBi’s electron current is >2 orders of magnitude lower due to trapping, while 3CzTrz is nearly trap-free at similar thickness (100 nm).
• Energy-level trends alone cannot explain transport differences: amorphous-phase DOS predicts similar trapping for fluorinated series, conflicting with experiments. Crystalline-phase DOS shows oxygen EAs are significantly lowered (deeper traps) in crystalline 1CzTrz-F and DTPT-D2FCz relative to amorphous, but still cannot account for trap-free transport in 3CzTrz-F and DTPT-DCz.
• XRD reveals packing-dependent shielding: 3CzTrz-F shows inclined face-to-face stacking of phenyl-substituted triazine cores forming a one-dimensional double layer along a crystallographic axis, with carbazole donors crowding and shielding the triazine acceptor core from small molecules ('closed' geometry). 1CzTrz-F exhibits more 'open' stacking with less shielding. DTPT-DCz exhibits edge-on carbazoles (tilt ~40°) and tert-butyl groups protecting triazine rings; DTPT-D2FCz shows π-stacking of fluorinated carbazoles that leaves triazine cores exposed.
• MAS SS-NMR detects symmetry-breaking and multiple 19F environments in both 1CzTrz-F and 3CzTrz-F thin films, indicating local molecular ordering rather than fully amorphous films.
• Electronic coupling analysis shows stronger O2–host coupling (and thus more effective/deeper traps) in crystalline 1CzTrz-F compared to 3CzTrz-F.
• Conclusion: Spatial separation of HOMO/LUMO on donor/acceptor parts plus molecular packing that shields the LUMO-bearing acceptor from impurities can effectively eliminate extrinsic electron trapping, expanding the practical trap-free window beyond the nominal 2.5 eV energetic criterion and enabling balanced ambipolar transport in wide band gap materials.

The study addresses whether intrinsic, simultaneous trap-free transport of electrons and holes can be realized in wide-band-gap organic semiconductors where at least one frontier orbital lies outside the universal trap-free energy window. The findings demonstrate that energetics alone (IE/EA values) do not determine trap susceptibility; instead, molecular packing that spatially protects the acceptor-localized LUMO from oxygen and water is crucial. Donor–acceptor architectures with three carbazole donors around a triazine acceptor create 'closed' stacking geometries that shield the electron-transport pathway, drastically reducing or eliminating effective electron traps despite EAs below 3.5 eV. Conversely, 'open' geometries expose the acceptor core, increasing oxygen coupling and trap depth. This packing-enabled shielding yields nearly trap-free, balanced ambipolar transport in 3CzTrz, with electron transport surpassing that of standard ETL TPBi. The generality is confirmed in a second series (DTPT-DCz vs fluorinated analogues): non-fluorinated DTPT-DCz attains near trap-free transport due to protective packing, while increased fluorination promotes packing that exposes the triazine core and increases trapping, contrary to expectations from EA trends. These results suggest that the traditional 2.5 eV window can be effectively broadened via molecular design that decouples impurity interactions from the LUMO by spatial shielding, offering a route to balanced transport in blue-emitting materials used in OLEDs.

By engineering donor–acceptor molecules so that HOMO and LUMO reside on different moieties and by tuning stacking to crowd the LUMO-bearing acceptor with donor groups, the authors eliminate extrinsic electron trapping and achieve nearly trap-free, balanced ambipolar transport in wide band gap organic semiconductors. The approach is validated across two molecular series (CzTrz and DTPT), where optimal 'closed' packing geometries lead to orders-of-magnitude higher electron currents and, in 3CzTrz, balanced trap-free electron and hole transport. This molecular design strategy effectively broadens the functional trap-free window beyond the nominal energetic limits and enables materials that outperform standard ETLs like TPBi. The work paves the way for efficient, potentially printed, single-layer blue OLEDs by mitigating trapping without relying solely on energy level alignment. Future research could: (i) systematically control and probe thin-film packing to maximize shielding in device-relevant morphologies; (ii) quantify impurity interactions (O2/H2O) in situ; (iii) expand the design to other donor–acceptor systems and polymers; and (iv) integrate these materials in operational devices to correlate packing, transport, efficiency, and stability.

• The relation between processing conditions and trapping remains under debate; while local ordering is evidenced by MAS SS-NMR, the precise control and quantification of thin-film packing versus single-crystal structures is limited.
• Computational DOS and trapping analyses for amorphous and crystalline phases capture energetic trends but do not fully explain trap-free transport in some cases, indicating limitations in modeling packing-related shielding and impurity accessibility.
• Single-crystal XRD structures inform plausible packing motifs, but films may contain mixed phases; direct visualization of impurity access pathways in thin films is not provided.
• Impurity types and concentrations (O2, H2O) are inferred from prior knowledge; direct in situ measurements of impurity uptake and their spatial distribution in operating devices are not reported.