Regular Tandem QLED Outcoupling

A tandem QLED stacks two emission units in one optical cavity. The two QD layers sit at different distances from the outcoupling interface and the metal cathode, so their outcoupling efficiencies are not equal and they respond differently to the same transport-layer thickness.

This case builds three models — bottom unit only, top unit only, and both — and sweeps the bottom-unit ZnMgO thickness from 22 nm to 200 nm to reproduce the three thickness responses of paper Figure 2i-k.

At 120 nm, the paper marks Air fractions of 22.4% for the bottom unit, 32.2% for the top unit, and 54.6% for their summed tandem contribution. The broad maxima and the larger top-unit contribution are the main comparison features.

Background

A tandem QLED places two emissive units in one optical stack. Under ideal charge balance, one injected electron can generate at most one exciton in each unit. Optically, the two QD layers occupy different positions relative to the output interface and metal cathode, so their outcoupling efficiencies are not equal and respond differently to the same transport-layer thickness.

The paper uses an ultrathin ITO charge-generation layer and reports an experimental EQE above 51%. This reproduction addresses the optical power fractions only. Charge generation, carrier balance, and lifetime are outside the model.

Structure

All three models share one stack and differ only in which QD layers have Emission enabled. The stack runs from the Glass output side to the Al cathode.

Once entered, the structure page looks like this. The two QD layers sit on opposite sides of the interconnect and share the Glass output channel:

Optical and Emitter Settings

Unit White here means the single-wavelength calculation uses no additional PL weighting; it does not indicate white device emission.

The bottom-unit emitter, enabled only in the bottom-unit model:

The top-unit emitter, enabled only in the top-unit model; the tandem model enables both:

Enable Mode only, with the wavelength set to Single at 632 nm:

The sweep variable is the bottom-unit ZnMgO thickness. All three models use the same 90 thickness values, so their curves are directly comparable:

On the results page, Stack shows the complete normalized power distribution across all seven channels for comparison with the paper's stacked-area plots, and Line retains only Top Outcoupling for reading the 120 nm design point and curve maxima. The seven channels map onto the paper's as follows.

Dreapex TMM additionally separates Bottom Outcoupling and Non-radiative. The former is nearly zero in this setup, and the latter is zero because the quantum efficiency is 1.

Simulation Results and Comparison with Figure 2

The paper separates the bottom, top, and summed tandem power fractions, so the simulation results follow the same i-j-k order. Each subsection first uses Stack to compare the complete modal distribution, then uses Line to read the Top Outcoupling values.

Figure 2i: bottom unit

Paper Figure 2i (bottom unit) is in the original article.

The Stack result of the same sweep, giving the full channel split versus thickness:

The seven channels close to 100% at all 90 thickness values. As ZnMgO thickens, the Evanescent band contracts rapidly while Waveguide, Substrate, and Top Outcoupling expand, reproducing the principal mode-transfer direction in Figure 2i.

Switching to Line leaves only Top Outcoupling, which is where the design-point value is read:

The bottom unit gives 17.35% at 120 nm and reaches 19.07% near 152 nm; the paper reports about 22.4% at 120 nm. The broad cavity optimum is reproduced, with visible offsets in both position and magnitude.

Figure 2j: top unit

Paper Figure 2j (top unit) is in the original article.

Switching to the model with only the top QD enabled and repeating the same sweep:

The top-unit Stack also closes to 100% at every thickness and retains a broad Top Outcoupling band. The internal high-in-plane-wave-vector power is partitioned differently between Waveguide and Evanescent than in the paper, so the SPP channel is not fitted point by point.

The same Line view:

The top unit gives 29.49% at 120 nm and peaks at 29.98% near 98 nm. Its optimum differs from the bottom unit's 152 nm, showing that the two QD layers occupy different optical positions. The contribution ordering at the design point agrees with the paper's 32.2% > 22.4%.

Figure 2k: tandem device

Paper Figure 2k (tandem device) is in the original article.

With both QD layers enabled at equal weight:

This Stack is the normalized average of two equal-weight emitters, so its total height is fixed at 1. Paper Figure 2k sums the two isolated units and therefore has a total height of 2. The screenshot compares relative channel composition and thickness trends; the quantitative Figure 2k comparison still uses the sum of the isolated-unit results.

The tandem Top Outcoupling curve:

The trace reaches 23.68% near 136 nm, gives 23.42% at the paper's 120 nm design, and remains between the two isolated traces.

The Figure 2k comparison uses the isolated-unit sum: 17.35% + 29.49% = 46.84%, which is 7.76 percentage points below the paper's 54.6%. At 120 nm, the dual-emitter 23.42% agrees with the arithmetic mean of the isolated results and is excluded from the 54.6% error calculation.

Deviation Notes

The 2 nm Al–Al₂O₃ interconnect uses the paper's single effective value 1.37 + 7.52i, and the 2PACz:6PA layer uses the scalar 1.72 surrogate reported for the optical calculation. Treating those two ultrathin mixed films as homogeneous changes both the bottom-unit loss and the optimum position.

This case computes at a single wavelength of 632 nm while the published curves are given across the full emission spectrum. That difference has been measured: switching the Mode wavelength mode to a 560–700 nm sweep and averaging flat over 600–660 nm at the 120 nm design point moves the bottom unit from 17.34% to 17.37% and the top unit from 29.49% to 29.23%. The spectral weighting is worth less than 0.3 percentage point and is not the source of the gap to the paper.

Each QD layer uses a central isotropic delta emitter; different recombination-zone positions or horizontal-dipole fractions would affect the two units differently. The model excludes electrical charge balance, so the experimental 51.2% EQE is not what this case compares against.

Back to Case Studies · Next: Tandem QLED Cavity Optimization