Top-Emitting QLED Outcoupling

This case reproduces the optical model in Figure 1b-d of Li et al.: the 626 nm mode partition of bottom-emitting and top-emitting red QLEDs, their power-dissipation spectra, and a two-dimensional sweep of the two IZO phase-tuning layers. The baseline Air fractions agree within 0.42 percentage point, and the sweep optimum falls one 5 nm grid step from the paper's top-IZO design.

Panel b shows that top emission removes the Substrate channel while slightly increasing the Air fraction; panel c resolves the redistributed optical modes, and panel d locates the high-efficiency cavity region near 150/100 nm bottom/top IZO.

Background

A bottom-emitting QLED loses part of its generated power into the thick glass substrate. A top-emitting device replaces that output path with a semitransparent electrode above a metal mirror, eliminating the substrate channel but introducing a strong microcavity. The IZO layers then act as phase-tuning layers: their thicknesses control how much emission couples to air, guided modes, absorption, and evanescent channels.

The paper also reports a 44.5% experimental EQE after adding an external scattering layer. This planar reproduction tests the calculated cavity and mode partition only; it does not model that scattering layer.

Mapping the Paper to the TMM Model

Bottom-emitting baseline (output side → Ag)

Top-emitting baseline (output side → Ag)

Reproduction Target and Acceptance Criteria

1. The bottom-emitting top-outcoupling fraction should be within one percentage point of the paper's 41.42% Air fraction.
2. The top-emitting top-outcoupling fraction should be within one percentage point of the paper's 45.74% Air fraction.
3. The substrate channel should disappear in the top-emitting model, while the waveguided fraction increases.
4. The two-dimensional sweep should place the strongest region near 150 nm bottom IZO and 100 nm top IZO.
5. The flat model is not expected to reproduce the 44.5% experimental EQE of the device with the external scattering layer.

Modeling Path in Dreapex TMM

Structure

Use the two baselines listed above; the IZO-thickness sweep starts from the top-emitting model.

Optical settings

Use a single 626 nm wavelength for Figure 1b-c. Enable Mode to obtain the integrated channel fractions and Power Dissipation to inspect their in-plane-wave-vector distribution. For Figure 1d, switch Mode to the spectrum-weighted 570–670 nm range.

Example Setup

Bottom-emitting model

Glass forms the Substrate channel unique to the bottom-emitting device; the top-emitting comparison has no such channel.

Both baselines share this emitter setting, so their mode-fraction difference comes from the stack alone.

This page produces the integrated channel results compared with Figure 1b.

It shares the 626 nm input with Mode, so Figure 1c changes only the result representation.

Top-emitting model and sweep

With the Glass channel removed, the top-emitting baseline should have zero Substrate fraction.

Figure 1d uses the PL-weighted result rather than the baseline's single-wavelength basis.

Each heatmap pixel represents one top/bottom IZO thickness pair; the sweep contains 1,600 pairs.

Simulation Results and Comparison with Figure 1

Panels b-d provide the paper values for the channel fractions, power-dissipation curves, and the IZO thickness landscape used in the three comparisons below.

Figure 1b: mode partition

The bottom-emitting Substrate channel is power that enters the 1 mm glass without escaping into air, so it remains separate from the Air channel in Figure 1b.

The top-emitting stack reduces Substrate to zero; most of the redistributed power increases Waveguide, while Top Outcoupling rises only slightly.

The paper reports Air fractions of 41.42% and 45.74%. The differences are +0.11 and −0.42 percentage point, respectively.

Figure 1c: power dissipation

The bottom-emitting curves establish the TE, TM, and Total peak shapes. Mode-channel boundaries are not marked, so the exact channel fractions come from the Mode result.

The top-emitting cavity changes the peak positions, widths, and area distribution. The two plots scale their vertical axes independently, and the exact channel fractions remain those reported by Mode.

Figure 1d: IZO thickness sweep

The paper design and the simulated maximum lie in the same broad high-value lobe, showing that the optimum is insensitive to one 5 nm top-IZO grid step.

The peak metric differs by 0.52 percentage point.

Deviation Analysis

1. Refractive-index and extinction-coefficient curves were digitized from the Supporting Information. Digitization spacing and interpolation shift the cavity phase slightly.
2. The two-dimensional search uses a 5 nm grid; the reported optimum can therefore move by one grid point even when the same broad maximum is reproduced.
3. The emitter is represented as one isotropic delta sheet in the center of the QD layer. A measured dipole-orientation ratio or recombination-zone profile could change the channel fractions.
4. The model is planar. The external scattering layer used by the highest-EQE experimental device is intentionally absent, as are roughness and lateral scattering.

Further Extensions

1. Refine the sweep to 1 nm steps around bottom IZO 140–160 nm and top IZO 90–115 nm.
2. Sweep dipole orientation to determine how much of the waveguided fraction can be recovered without changing the cavity.
3. Add a second sweep over QD position or ZnMgO thickness to quantify fabrication tolerance.
4. Compare the Mode result at 626 nm with the spectrum-weighted result to separate peak-wavelength and broadband design choices.

Back to Case Studies · Next: Regular Tandem QLED Outcoupling