Dipole Orientation and Angle-Resolved Photoluminescence of TADF Doped Films

Crovini et al. (2024): p-polarized angle-resolved photoluminescence and dipole orientation of five DMAC-TRZ derivative co-evaporated films

AuthorLuke Cole

Aryl-Substituted Acridine Donor Derivatives Modulate the Transition Dipole Moment Orientation and Exciton Harvesting Properties of Donor–Acceptor TADF Emitters

Authors: Ettore Crovini, Kleitos Stavrou, Prakhar Sahay, et al. (corresponding authors Wolfgang Brütting, Andrew Monkman, and Eli Zysman-Colman)

Journal: The Journal of Physical Chemistry C 128, 14429 (2024) · Comparison target: Figure 6

Swap one donor substituent on a TADF molecule and the main lobe of the p-polarized angle-resolved photoluminescence of the co-evaporated film rises from about 0.73 to about 0.95, corresponding to a vertical dipole fraction going from 0.15 to 0.26.

This case builds about the simplest model there is: one 35 nm doped film between fused silica and air. The five published orientations go in as given, with no fitting, and five angle sweeps reproduce the five curves of paper Figure 6.

Crovini Figure 6 — p-polarized angle-resolved photoluminescence of five DMAC-TRZ derivatives at 10 wt% in co-evaporated films; squares are measurements and the solid lines are the authors' own dipole-model fits.Crovini et al., The Journal of Physical Chemistry C 128, 14429 (2024), Figure 6 CC BY

All five curves sit near the normalized peak around 0°, drop into a sharp minimum near 43°, form a main lobe at 55°–57°, and then fall monotonically to 90°. A taller main lobe means a larger vertical dipole fraction.

Background

Angle-resolved photoluminescence couples the doped film to a half-cylinder prism and records the s- and p-polarized emission intensity angle by angle on a rotation stage. Dipoles perpendicular to the film send energy to large angles and dipoles parallel to the film concentrate it at small angles, so the normalized p-polarized distribution can be inverted for the orientation.

The sharp minimum near 43° comes from total internal reflection at the film–air interface: the critical angle for fused silica at $n=1.4623$ is about 43°, beyond which the air side becomes evanescent and the interference condition changes abruptly. All five curves kink at the same angle.

Orientation appears in three conventions in the literature: vertical fraction

a

, horizontal fraction

1-a

, and order parameter

S=(3a-1)/2

. This paper reports the vertical fraction, the same convention as the Vertical Fraction field, so the values go in unchanged. Entering

1-a

by mistake degrades the all-angle RMS of the five curves from 2.46 %–4.68 % to 17.88 %–47.60 %.

Structure

The samples are organic films co-evaporated on glass and index-matched to a fused-silica half-cylinder prism. The matching fluid removes the back-surface interface, and the half-cylinder lets emitted light cross the outer prism surface along the normal, so the stage reading equals the emission angle inside the prism and the detection side is treated as semi-infinite fused silica.

Position	Material	Thickness	Optical input
`Incidence Medium` (detection side)	Fused silica	Semi-infinite	$n=1.4623$ , $k=0$
Emissive layer	10 wt% derivative co-evaporated in mCBPCN	`35 nm`	Per-model values at each fitting wavelength, below
`Transmission Medium`	Air	Semi-infinite	$n=1$ , $k=0$

The emissive layer takes the measured optical constants of the neat mCBP-CN host, ignoring the correction from 10 wt% doping. Each model needs only one pair of values, at its own fitting wavelength.

Fitting wavelength	$n$	$k$
`490 nm`	1.7518	0.00353
`500 nm`	1.7468	0.00329
`520 nm`	1.7381	0.00289
`521 nm`	1.7377	0.00287
`539 nm`	1.7310	0.00260

The structure table holds a single finite-thickness film, bounded by fused silica and air:

Structure page of the tBuPh-DMAC-TRZ model with a fused-silica incidence medium, a 35 nm emissive layer, and an air transmission medium

The paper does not give the doped-film thickness. 35 nm is the only free structural parameter in this case, fixed by requiring one thickness to fit all five curves: fitting each curve independently gives scattered values of 28–39 nm, but the five confidence bands are simultaneously consistent only at 35 nm.

Optical and Emitter Settings

Only three values differ between the five models: the emissive-layer $n$ and $k$ , and the emitter's Vertical Fraction.

Model	Fitting wavelength	`Vertical Fraction`	Emissive-layer $n$	Emissive-layer $k$
dPh-DMAC-TRZ	`520 nm`	`0.15`	1.7381	0.00289
CNPh-DMAC-TRZ	`490 nm`	`0.16`	1.7518	0.00353
OMePh-DMAC-TRZ	`539 nm`	`0.20`	1.7310	0.00260
tBuPh-DMAC-TRZ	`521 nm`	`0.25`	1.7377	0.00287
CF3Ph-DMAC-TRZ	`500 nm`	`0.26`	1.7468	0.00329

Everything else is shared:

Setting	Value
Spectrum type	`Unit White`
Dipole orientation	`Custom`, `Vertical Fraction` set to the published vertical fraction
Dipole position and distribution	Relative position `0.5`, `Delta`
Detector	`Intensity`
Wavelength mode	`Single`, at each emitter's fitting wavelength
Angle mode	`Sweep`, `0°–89.5°`, step `0.5°`
Comparison quantity	The `TM` curve of `Normalized Angular Distribution`

The emitter panel should show a Unit White spectrum, Custom orientation, the matching Vertical Fraction, relative position 0.5, and a Delta distribution:

$Emitter settings for tBuPh-DMAC-TRZ with Custom orientation and a vertical fraction of 0.25$

Angles are defined inside the Incidence Medium, the same quantity as the paper's stage reading:

Intensity detector configured for a single wavelength of 521 nm and an angle sweep from 0 to 89.5 degrees

A single-wavelength comparison needs no measured photoluminescence spectrum: the normalized angular distribution is evaluated at a fixed wavelength, so the spectral amplitude cancels in the normalization.

Before normalization the angular distribution is linear in the orientation value. To compare a batch of orientations, run Parallel and Perpendicular once each and combine them linearly by fraction, instead of rerunning every Custom value.

Simulation Results and Comparison with Figure 6

The published figure, with squares for measurements and solid lines for the authors' own dipole-model fits:

The five computed curves overlaid on the digitized measurements:

Six-panel figure with the paper's measured points, the paper's fits, and the Dreapex TMM curves for five emitters, plus a panel comparing recovered and reported orientations — Measured points from paper Figure 6, the authors' own fits, and the Dreapex TMM results overlaid. The bottom-right panel compares the recovered orientations with the reported values, with horizontal bars for the published uncertainties.Paper data from Crovini et al. Figure 6 and Table 2; Dreapex TMM data from the real runs of this case; independently plotted CC BY 4.0 (independent plot)

Emitter	Fitting wavelength	Published vertical fraction	Digitized points	All-angle RMS
dPh-DMAC-TRZ	`520 nm`	0.15 ± 0.010	29	3.92 %
CNPh-DMAC-TRZ	`490 nm`	0.16 ± 0.019	18	3.99 %
OMePh-DMAC-TRZ	`539 nm`	0.20 ± 0.022	43	4.68 %
tBuPh-DMAC-TRZ	`521 nm`	0.25 ± 0.010	66	3.75 %
CF3Ph-DMAC-TRZ	`500 nm`	0.26 ± 0.008	11	2.46 %
All five combined	—	—	167	3.83 %

For reference, the measured squares scatter 1.46 % to 4.76 % around the paper's own fitted lines in the same figure — the same order as the model residual, and the digitization noise floor of Figure 6.

The five raw results from the app, ordered by increasing vertical fraction. dPh has the lowest orientation and the shortest main lobe:

Normalized angular distribution of dPh-DMAC-TRZ at 520 nm

CNPh at 0.16:

Normalized angular distribution of CNPh-DMAC-TRZ at 490 nm

OMePh at 0.20:

Normalized angular distribution of OMePh-DMAC-TRZ at 539 nm

tBuPh at 0.25:

Normalized angular distribution of tBuPh-DMAC-TRZ at 521 nm

CF3Ph has the highest orientation and the tallest main lobe:

Normalized angular distribution of CF3Ph-DMAC-TRZ at 500 nm

Recovering orientation from the curves

Releasing the orientation while keeping the shared 35 nm thickness and fitting each curve separately is how the paper extracts orientation in the first place.

Emitter	Published	Recovered here	Difference	Published uncertainty	RMS after recovery
dPh-DMAC-TRZ	0.15	0.131	−0.019	±0.010	3.73 %
CNPh-DMAC-TRZ	0.16	0.147	−0.013	±0.019	3.79 %
OMePh-DMAC-TRZ	0.20	0.199	−0.001	±0.022	4.68 %
tBuPh-DMAC-TRZ	0.25	0.236	−0.014	±0.010	3.54 %
CF3Ph-DMAC-TRZ	0.26	0.243	−0.017	±0.008	2.31 %

The ordering matches exactly: 0.131 < 0.147 < 0.199 < 0.236 < 0.243 against the published 0.15 < 0.16 < 0.20 < 0.25 < 0.26. The orientation span is 0.112 here and 0.110 in the paper.

All five differences share a sign, with a mean of −0.013, so this is a systematic downward shift. The confidence interval on a single curve is about ±0.02, so the workflow separates 0.15 from 0.26 but not 0.15 from 0.16.

Deviation Notes

Three inputs are not published, and together they account for the −0.013 systematic shift: the film thickness (taken as 35 nm, with no independent measurement to check against), the doped-film $n$ and $k$ (taken from the neat host, ignoring the 10 wt% doping correction), and the detection-medium index (fused silica at 1.4623 for 500 nm, with no model for residual mismatch between the matching fluid and the glass). Lowering that index from 1.4623 to 1.45 raises the recovered orientation by 0.017; replacing the point emission depth with a uniform in-layer distribution lowers it by 0.009.

The reference values come from independent digitization of Figure 6, not from numerical source data. The paper's own fitted lines are unreliable between 42° and 46° and above 75°, so the main comparison uses only the measured squares.

Back to Case Studies