Tandem QLED Cavity Optimization

The optimum cavity length of a top-emitting QLED is not a point but a diagonal band: a slightly thicker electron-transport layer and a slightly thinner top electrode restore the same phase. A tandem device has to make that band serve two emission units at different positions.

This case runs one two-dimensional thickness sweep for a single-emitting-layer device and one for a tandem device, using spectrum-weighted Mode to produce the two cavity maps of paper Figure 2c-d.

Panels a-b identify the two cavities; panel c places the single-layer maximum at 36.56%, while panel d places the tandem maximum at 31.87% and shifts the high-value region to the top-unit cavity dimensions.

Background

Top-emitting QLEDs form a microcavity between a semitransparent IZO electrode and an Ag mirror. Changing either the electron-transport-layer thickness or the top-electrode thickness changes the optical phase, field distribution, and fraction of generated light that reaches air. A tandem device adds a second emitting unit, so a useful cavity must support both source positions across the QD emission band.

The paper reports both optical optimization and device stability. This case reproduces only the optical maps EQE/γ; it does not model drive voltage, interconnecting-layer charge generation, or lifetime.

Structure

Both stacks run from the output side to the Ag mirror. The 100 nm Ag is optically thick, so glass behind it is omitted from the calculation.

Single-emitting-layer device

Tandem device

Downloadable refractive-index data

The five .nk files below were extracted from Figure S5 of the paper's supporting information. The RQD and TCTA files contain $n$ only, so Dreapex TMM imports their missing $k$ values as zero. Use the TCTA file for MoO₃ as well, following the approximation stated in the supporting information.

The single-emitting-layer structure page, with the two swept layers on opposite sides of the QD source jointly tuning the cavity phase:

The tandem structure places one QD source on each side of the interconnect, at different cavity positions:

Optical and Emitter Settings

Both red-QD layers use the same extracted PL spectrum and the same emitter settings; the tandem model only adds a second source position.

Both emitter layers use Spectrum-acs.nanolett.4c02021.pl, extracted from the red-QD PL spectrum in Figure S1. These datasets are provided for reproducing this case; cite Yuan et al. (2024) when reusing them.

Both QD layers in the tandem model reuse the same settings, so the top-unit panel looks identical:

Weighted Average includes the QD PL weight in every thickness combination's Mode result:

The two sweep variables for the single device:

The tandem device sweeps the two layers of the top unit instead:

The paper combines the two active units as an equal-weight average, which the two-emitter setting at weight 1 implements directly. With q₀ = 0.9 and the paper's convention of dividing out the charge-balance factor, Mode Top Outcoupling maps to "effective radiative yield × top-outcoupled fraction" and serves as the optical $EQE/\gamma$ quantity. It is neither a pure light-extraction efficiency nor a measured EQE, because the device's actual charge balance is not included.

Simulation Results and Comparison with Figure 2

Figure 2c: single-emitting-layer cavity map

Paper Figure 2c is in the original article.

The same kind of heatmap after all 4,104 thickness pairs complete:

High values form curved bands across both thickness axes, showing that ZnMgO and top IZO compensate cavity phase together. All 4,104 points completed without errors. The maximum is 35.80% at 75 nm ZnMgO and 110 nm top IZO. The paper's strongest region is centered on the same nominal cavity and reaches 36.56%, a difference of 0.76 percentage point.

Figure 2d: tandem cavity map

Paper Figure 2d is in the original article.

The tandem device across its 2,916 combinations:

This reproduction uses 4 nm steps for top ZnMgO and 2 nm steps for top IZO so the complete result remains renderable in the browser chart. The main ridge is centered around 112–120 nm top ZnMgO and 100–106 nm top IZO and extends diagonally along both axes, showing that the two thicknesses compensate cavity phase together. All 2,916 points completed without errors. The maximum is 31.87% at 112/106 nm; the paper's nominal 120/100 nm design gives 31.72%, only 0.14 percentage point below the maximum.

The single-device result closely reproduces the peak magnitude and design region. The tandem peak magnitude is effectively identical on the reported precision; its optimum sits 8 nm lower in top ZnMgO and 6 nm higher in top IZO than the nominal paper design.

Deviation Notes

The supporting information gives refractive indices for RQD and TCTA but no extinction coefficients, so k is explicitly set to zero for those layers, and MoO₃ reuses the TCTA constants following the same approximation. Both the PL and the optical-constant curves come from digitization, so sampling and interpolation can move a narrow cavity optimum by a few nanometres.

Each EML uses a central isotropic Delta emitter; a distributed recombination zone or unequal unit weights would change the tandem map. The planar model contains no electrode roughness or scattering, and predicts the optical $EQE/\gamma$ map rather than a measured device EQE or lifetime.

Back to Case Studies