Electrical excitation of insulators at the nanoscale was until yesterday an almost impossible mission: materials with a large bandgap require extreme conditions, so researchers mainly relied on optical excitation. However, a new generation of lanthanide-doped insulating nanoparticles (LnNPs) is now getting a completely different “switch”: molecular triplets on organic ligands that transfer energy to lanthanide ions and thus turn on very pure, narrowband emission in the second near-infrared window (NIR-II, 1000–1700 nm). The key progress is the demonstration of electroluminescence (EL) from these insulating systems at low voltages, opening the way toward hybrid LEDs, and in the long term toward electrically pumped lasers for biomedicine, optogenetics, and communications.

Why lanthanide nanoparticles are special – and why they were “unswitchable”

LnNPs consist of a dielectric, insulating host (typically fluorides like NaGdF₄, NaYF₄ or NaLuF₄) with embedded lanthanide ions (Nd³⁺, Yb³⁺, Er³⁺ etc.). Such a system offers exceptional chemical and photostability as well as very narrow and tunable emission lines in the NIR-II region, unlike organic dyes and colloidal quantum dots whose spectra are broad due to homogeneous broadening. Precisely this narrowband nature and stability make LnNPs attractive for deep imaging solutions, precise spectroscopy, and optical links with high channel density.

The problem is that LnNPs are – insulators. A wide bandgap (~8 eV) prevents standard electrical charge injection as in semiconductors, so until recently there was no way to directly derive EL from them under low voltages. While quantum dots, perovskites, or organic semiconductors routinely work in LED structures, LnNPs, without additional tricks, remained “optics-only”.

Triplet-mediated excitation: organic ligands as antennas

The key turnaround comes from combining LnNPs with selected organic molecules that generate triplet excitations and transfer energy to the f-f transitions of lanthanides. In practice, researchers partially replace the usual oleic acid (OA) on the nanoparticle surface with the ligand 9-anthracenecarboxylic acid (9-ACA). On the organic part, after electrical injection, singlets and triplets are formed in a 1:3 ratio (spin statistics). The 9-ACA triplets then transfer energy to the ionic levels of Nd³⁺, Yb³⁺ and Er³⁺ via the mechanism of Dexter transfer (TET, triplet energy transfer) at very short distances. Since the triplet on 9-ACA is long-lived (hundreds of microseconds), and TET takes place in microseconds, the transfer is highly efficient, with few competing processes.

Experimentally, binding 9-ACA to LnNPs results in strong UV absorption inherited from the molecule and multiply enhanced NIR-II photoluminescence (e.g., ~6.6× for Nd, ~34× for Yb, ~24× for Er under 350 nm excitation). Furthermore, FTIR spectroscopy and DFT simulations indicate that 9-ACA preferentially coordinates to Ln³⁺ sites (unlike OA which also binds to Na⁺), which further favors ligand–ion energy coupling.

First LnLEDs: structure, operation, and spectral features

What does the device look like? A typical architecture is glass/ITO/PEDOT:PSS/poly-TPD/LnNP@9-ACA/TmPyPB/LiF/Al. ITO and LiF/Al are electrodes; PEDOT:PSS facilitates hole injection, poly-TPD and TmPyPB serve as HTL and ETL, and the hybrid LnNP@9-ACA forms the emissive layer. Electrons and holes recombine primarily on 9-ACA, creating singlets and triplets; the latter transfer energy with very high efficiency to lanthanide ionic levels, which then emit in the NIR-II region.

Measured EL shows extremely narrow bands: for example, full widths at half maximum (FWHM) are approximately 20 nm (Nd), 43 nm (Yb), and 55 nm (Er) – multiple times narrower than typical NIR systems based on quantum dots (often >150 nm). Peak wavelengths are centered around ~1058 nm (Nd), ~976 nm (Yb), and ~1533 nm (Er), with no shift with voltage. Devices turn on already at approximately 5 V, and withstand higher voltages (>15 V), which is attributed to the fact that most energy-rich triplets in the organics are “drained” to robust lanthanide 4f levels, thereby mitigating degradation typical for the organic component.

Evidence for triplet transfer: dynamics and oxygen as a “quencher”

Kinetic experiments further confirm the scenario. Bound 9-ACA shows a shortened singlet lifetime (from ~12.4 ns to <5 ns, depending on Ln), indicating accelerated intersystem crossing (ISC) induced by the proximity of ions with unpaired spin. Transient absorption measures rapid growth of the triplet signal (≈1.4–1.9 ns) and its significantly faster decay than in pure 9-ACA, which is a direct consequence of TET to levels 2F_5/2 (Yb), 4F_3/2 (Nd) and 4I_11/2 (Er). Under air conditions, photoluminescence in the NIR is strongly quenched due to oxygen (≈50% and more), which is a typical signature of triplet quenching and additional confirmation that TET is the dominant channel.

From photons to current: the difference between photo- and electro-excitation

Comparing PL and EL spectra reveals that ratios of individual bands differ, implying different energy transfer pathways under photo- and electro-excitation. In electrical excitation, recombination occurs on 9-ACA in a thin organic-inorganic monomolecular layer, so surface ions may have a preferential role in accepting triplet energy. This also explains traces of visible EL: the blue component originates in poly-TPD, and the red from the HTL/ETL interface at places where the hybrid layer is not ideally homogeneous – which are loss channels and optimization targets for the next generation of devices.

How to increase efficiency: core–shell, leakage blocking, and light extraction

The first LnLEDs achieve NIR EQE in the range of approximately 0.004–0.04% (Er–Yb), and by using a core–shell design Yb@Nd, optimized transport layers with stronger electron/hole blocking and microlenses for light extraction, peak EQE exceeds ~0.6% in the NIR-II region. Although this is still far from the best QD LEDs in the visible, one must bear in mind that this involves an insulating host with rare-earth ions and that this is the first generation of devices. Main hurdles are: (i) limited PLQE of highly doped cores without a shell, (ii) charge leakage through the emissive monolayer and recombination in transport layers, and (iii) poor NIR-II light extraction from the thin film.

Improvement strategies include dedicated ligand design with higher OA replacement (currently <10% sites covered by 9-ACA depending on ionic species), multilayer emitter structures instead of a monolayer, enhanced electron/hole blocking, and use of optical extraction elements (microlenses, textured substrates). At the emitter level, increasing PLQE through optimization of dopant concentration, particle type and size will be crucial; literature already knows examples of Er³⁺ with PLQE >50% at 1530 nm, suggesting real room for EQE growth.

Where LnLEDs excel: NIR-II as a window for tissues and data

The NIR-II window offers deeper penetration through biological tissues, lower scattering, and less autofluorescent noise. This makes sources with narrow lines and tunable peak position extremely desirable for non-invasive diagnostics, intraoperative navigation, photothermal therapy, and real-time pharmacokinetics monitoring. The high spectral purity of LnLEDs can reduce channel overlap in multichannel (multiplex) measurements, while material stability opens the way toward sterilizable, long-lasting probes and flexible surface emitters for wearable devices.

Another major area is optical communication. Narrower bands and electrical tuning of dopant combination allow “hopping” across channels with minimal crosstalk, while organic processing and low-voltage operation suggest potentially more favorable sources for short-range and integrated photonic links, including communication through turbid environments (fog, smoke, scattering media).

Comparison with competing platforms

• Quantum dots (QD): excellent efficiency and simple electrical addressing, but in the NIR-II region they mostly retain broad emissions (>150 nm FWHM), limiting spectral channel density and precise spectroscopy.


• Perovskites: high performance in visible and near IR, but stability and presence of lead remain challenges, especially for biomedical applications. Achieving very narrow lines in the NIR-II region is even harder.


• Organic emitters: easily processable, but triplet deactivations and broad bands often limit performance above 1000 nm.


• LnLED hybrids: currently lower EQE, but completely unique narrowband nature, emission tuning only by ion replacement, and potential for robust operation at higher voltages thanks to triplet “drain” to 4f levels.

Materials and nanoarchitecture: what is crucial

Efficient TET requires: (1) large spectroscopic overlap between ligand triplet phosphorescence (9-ACA spans approximately 1.3–1.9 eV) and Ln³⁺ absorption bands; (2) short donor–acceptor distance (achieved by 9-ACA chemisorption on Ln³⁺ sites); (3) sufficiently long triplet duration (≥100 µs), significantly longer than microsecond TET times; (4) inert environment or oxygen protection for photoluminescence measurements. Also, core–shell architectures (e.g., Yb@Nd) can separate absorption and emission sites, reduce back transfer and quenching, and increase PLQE.

Electronic device design: charge balance and blocking

To increase EL, it is necessary to balance electron and hole injection into 9-ACA, minimize leakage towards HTL, and suppress unwanted recombination in transport layers. Improved HTL with better hole injection and stronger electron blocking, and optimized ETL with appropriate work functions and level matching, are proven to raise EQE. Optimization of layer thickness and index contrast, in turn, increases NIR-II light extraction, which is an often overlooked loss in thin films.

The bigger picture: convergence of two worlds

Conceptually, LnLEDs combine the best of two worlds: organics that generate and control excitons and inorganic ionic centers whose 4f levels emit with atomically narrow lines. Precisely this convergence – triplet-mediated transfer towards 4f emitters – shows how “dark” states in organics become useful energy currency for turning on insulators. Comparatively, other teams are also demonstrating alternative approaches of electrogenerated excitons in lanthanide nanocrystals with emphasized spectral purity and tunability, confirming that NIR-II lanthanide sources are entering a new, electrically addressable phase.

From lab to application: what follows

To pave the way toward practical NIR-II sources for medical devices, endoscopic probes, or short-range optical links, it will be necessary to: (i) increase emitter PLQE (goal: tens of percent in NIR-II), (ii) shape thicker, more uniform active layers without voids to avoid visible EL and leakage, (iii) standardize ligands and replacement methods to achieve higher coverage than current <10% and (iv) integrate passive optical elements for efficient light extraction. Given the rapid progress of the field, combining core–shell design with new, triplet-friendly ligands and optimized transport layers emerges as the most likely path toward significantly higher EQE values and stability under continuous operation.

Note on date: data and results mentioned in this article were verified on December 6, 2025, with references to peer-reviewed papers and research communications published during November and December 2025.