Quantifying the Role of Higher-Lying Excited States in Organic Emitters via Multistate Ab Initio Kinetic Modeling

This paper introduces KinLuv, a comprehensive multistate *ab initio* kinetic modeling framework that explicitly incorporates higher-lying excited states and Herzberg-Teller vibronic coupling to quantitatively reproduce experimental photophysical observables and establish criteria for determining when simplified models are sufficient for designing high-performance organic emitters.

The Gist

Imagine you are trying to understand how a glowing firefly works. For a long time, scientists thought the process was simple: The firefly gets a burst of energy (like a spark), jumps up to a "high-energy" level, and then immediately jumps back down, releasing a flash of light.

This paper is about realizing that the firefly's journey is actually much more complicated than that simple up-and-down jump. Sometimes, the firefly doesn't just go straight back down; it gets stuck in a "waiting room" (a dark state) or even visits a second, higher "attic" before finally deciding to flash.

Here is a breakdown of the paper's discoveries using simple analogies:

1. The Old Map vs. The New Map

The Problem: For years, scientists used a "Three-State Map" to predict how organic light-emitting molecules (used in OLED screens) work. This map only looked at three places:

• The Ground Floor (S0): The molecule resting.
• The Main Floor (S1): The excited state where light is emitted.
• The Basement (T1): A dark, trapped state where the molecule loses energy without making light.

The Issue: This map worked well for some molecules, but for others, the predictions were wrong. It was like trying to navigate a city using a map that only showed the main street, ignoring the side alleys and highways that people actually use.

The Solution: The authors, Yue He and Daniel Escudero, created a new, more detailed map called KinLuv. This map includes "Higher-Lying States" (like an attic or a second basement). They realized that sometimes, molecules take a detour through these extra rooms, which changes how bright the light is and how long it lasts.

2. The "Vibronic" Elevator (The Secret Sauce)

One of the biggest discoveries in the paper is how the molecules move between these rooms.

• The Old Idea: Scientists thought the molecule just needed enough energy to jump from the Basement (T1) back to the Main Floor (S1). If the jump was too high, it wouldn't happen.
• The New Discovery: The authors found that the molecule is constantly vibrating (shaking) like a jelly. These vibrations act like an elevator or a boost.
  • Imagine trying to jump over a high fence. If you just run and jump, you might fail. But if you are on a trampoline (the vibration), you get a huge boost and clear the fence easily.
  • In the paper, this "trampoline effect" is called Herzberg-Teller (HT) coupling. They found that ignoring this vibration is like ignoring the trampoline; it leads to terrible predictions. When they included the "trampoline," their calculations finally matched real-world experiments.

3. Testing the Map on Different "Fireflies"

The team tested their new KinLuv map on six different types of glowing molecules. They found that the "right" map depends on the specific molecule:

• The "Simple" Fireflies (DOBNA & DABNA-1):

  • Analogy: These molecules are like people who take the elevator straight down. They don't really care about the attic or the basement.
  • Result: For these, the old "Three-State Map" was actually good enough. Adding the extra rooms didn't change the final result much.

• The "Complex" Fireflies (DiKTa & DQAO):

  • Analogy: These molecules are like busy commuters. They get stuck in the "Basement" (T1) for a long time. To get back to the Main Floor, they need to visit the "Attic" (T2) first to find a shortcut.
  • Result: If you use the old map, you get the brightness wrong. You must include the extra rooms (T2) to get the right answer.

• The "Tricky" Fireflies (DBT & PPDs-1):

  • Analogy: These are the troublemakers. They don't just visit the basement; they sprint up to the "Attic" (S2) and then fall straight to the ground without ever making light.
  • Result: The old map failed completely here. To understand why these molecules are so dim, you had to include the "Attic" (S2) in the model. Without it, the model predicted they would be bright, but in reality, they are dark.

4. Why This Matters

Why should you care about a molecule's "attic"?

• Better Screens: This research helps engineers design better OLED screens (like on your phone or TV). By knowing exactly which "rooms" the energy visits, they can design molecules that waste less energy and shine brighter.
• Saving Time and Money: Instead of building a super-complex model for every single molecule (which takes forever to calculate), this paper gives scientists a rule of thumb:
  • If the molecule is simple, use the 3-room map.
  • If the molecule is complex or has fast vibrations, use the 5-room map.
• The "Goldilocks" Zone: It teaches us when to keep things simple and when to get complicated, ensuring we get accurate results without doing unnecessary work.

The Bottom Line

This paper is like upgrading from a basic street map to a GPS with traffic updates. It shows us that to truly understand how organic lights work, we can't just look at the start and finish line. We have to watch the whole journey, including the detours, the vibrations, and the hidden rooms, to predict exactly how bright the light will be.

Technical Summary

1. Problem Statement

Traditional modeling of organic chromophore photophysics, particularly for Thermally Activated Delayed Fluorescence (TADF) materials, relies on a simplified three-state model involving only the ground state ($S_0$), the lowest singlet excited state ($S_1$), and the lowest triplet excited state ($T_1$).

• Limitations: This approximation assumes $S_1$ and $T_1$ are well-separated from higher-lying states ($S_{n>1}, T_{n>1}$). However, experimental and theoretical evidence increasingly suggests that higher-lying states play critical roles in:
  • Anti-Kasha emission.
  • Intersystem crossing (ISC) and reverse ISC (rISC) pathways mediated by "dark" intermediate states.
  • Internal conversion (IC) routes inaccessible from $S_1$.
• The Gap: While the qualitative importance of higher states is recognized, their explicit and quantitative impact on key observables (Photoluminescence Quantum Yield - PLQY, and prompt/delayed fluorescence lifetimes) has not been systematically assessed within a fully ab initio kinetic framework. Existing computational methods often neglect Herzberg-Teller (HT) vibronic coupling, leading to significant underestimations of (r)ISC rates (e.g., orders of magnitude errors in previous studies of DABNA-1).

2. Methodology

The authors developed a comprehensive computational framework combining high-level quantum chemistry with a custom kinetic modeling tool.

A. Computational Protocol

• Geometry & Energies: Ground and excited state geometries were optimized using TD(A)-CAM-B3LYP. To ensure accuracy in energy gaps (crucial for rate constants), single-point energy calculations were performed using Spin-Component Scaled Second-Order Algebraic Diagrammatic Construction [SCS-ADC(2)].
• Rate Constant Calculation: All radiative and non-radiative rate constants were computed using Fermi's Golden Rule (FGR) via the Thermal Vibration Correlation Function (TVCF) formalism.
  • Vibronic Coupling: Crucially, the calculations explicitly included Herzberg-Teller (HT) vibronic coupling effects, breaking the Condon approximation. This accounts for the dependence of transition dipole moments and spin-orbit coupling matrix elements (SOCMEs) on nuclear displacements.
  • Models: Adiabatic Hessian (AH) models with Duschinsky rotation were used.
• Software: Rate constants were calculated using standard packages (Gaussian, Turbomole, ORCA, Q-Chem, FCClasses).

B. Kinetic Modeling: KinLuv

The authors developed KinLuv, a Python-based tool to solve the kinetic equations.

• Models Tested: A hierarchy of kinetic models was constructed to test the necessity of higher states:
  1. Two-state: $S_0, S_1$
  2. Three-state: $S_0, S_1, T_1$ (Standard TADF model)
  3. Four-state: $S_0, S_1, T_1, T_2$ (Includes higher triplet)
  4. Five-state: $S_0, S_1, S_2, T_1, T_2$ (Includes higher singlet and triplet)
• Simulation: The tool solves coupled Ordinary Differential Equations (ODEs) for population dynamics over time (transient kinetics) and uses Steady-State Approximation (SSA) to derive PLQY.

C. Target Systems

Six molecules were selected to cover diverse photophysical regimes:

• MR-TADF Prototypes: DOBNA and DiKTa.
• Derivatives: DABNA-1 (DOBNA derivative) and DQAO (DiKTa derivative).
• Non-TADF Chromophores: Dibenzothiophene (DBT) and PPDs-1 (chosen for strong higher-state involvement).

3. Key Contributions

1. Development of KinLuv: A robust, open-source Python framework for multistate kinetic modeling that integrates fully ab initio rate constants.
2. Explicit HT Inclusion: Demonstrated that neglecting HT effects leads to massive underestimation of (r)ISC rates. The inclusion of HT is essential for quantitative accuracy, particularly for systems where direct $S_1 \leftrightarrow T_1$ coupling is symmetry-forbidden or weak.
3. Systematic Model Selection Criteria: Established clear physical criteria for determining when simplified (3-state) models are sufficient versus when complex (4- or 5-state) models are required.

4. Key Results

A. Mechanistic Insights into Rate Constants

• Vibronic Dominance: For the studied MR-TADF emitters, (r)ISC rates are dominated by HT contributions rather than Franck-Condon (FC) terms.
• Structural Influence: In DiKTa, slight molecular torsion allows out-of-plane vibrations that break symmetry, significantly enhancing SOCMEs via HT effects compared to the planar DOBNA. This explains DiKTa's faster ISC rates despite a slightly larger $S_1-T_1$ gap.

B. Performance of Kinetic Models

The study compared simulated PLQY and lifetimes against experimental data across the four model complexities:

• DOBNA & DABNA-1 (Slow (r)ISC regime):

  • The three-state model was sufficient to reproduce experimental PLQY and lifetimes.
  • Reasoning: (r)ISC processes are slow relative to fluorescence and IC. Higher-lying states ($T_2, S_2$) are not significantly populated, making their inclusion mechanistically interesting but quantitatively unnecessary for observables.

• DiKTa & DQAO (Fast (r)ISC regime):

  • The three-state model failed to predict PLQY accurately (overestimated).
  • The four-state model (including $T_2$) was essential to match experimental PLQY.
  • Reasoning: Fast ISC leads to substantial $T_1$ population. $T_2$ acts as a competing non-radiative decay pathway ($S_1 \to T_2 \to T_1 \to S_0$) that reduces PLQY. Ignoring $T_2$ misses this loss channel.

• DBT (Non-TADF, Fast ISC):

  • Similar to DiKTa, the four-state model was required. $T_2$ introduces a major non-radiative pathway ($S_1 \to T_2 \to T_1 \to S_0$), reducing PLQY from ~73% (2-state) to ~8.8% (4-state), matching the experimental 9%.

• PPDs-1 (Non-TADF, Fast rIC):

  • The five-state model (including $S_2$) was essential.
  • Reasoning: A highly efficient non-radiative pathway $S_1 \to S_2 \to S_0$ (mediated by fast rIC) dominates the dynamics. The 3-state model predicted a PLQY of ~95%, whereas the 5-state model correctly predicted ~2.9% (experimental 0.9%).

C. Accuracy and Limitations

• While the framework improves mechanistic understanding, absolute quantitative accuracy is still limited by uncertainties in ab initio rate constants (e.g., sensitivity to energy gaps).
• However, the study successfully identified that mechanistic completeness does not always equal quantitative improvement; the "minimal robust model" depends on the specific kinetic regime of the molecule.

5. Significance

• Rational Design: The framework provides a principled strategy for computational chemists to select the minimal kinetic model required for a specific molecule, balancing physical insight with numerical robustness.
• Beyond TADF: The methodology is applicable to a broad range of organic emitters, not just TADF, addressing the "Kasha's rule" violations and complex decay pathways in non-TADF systems.
• Validation of HT Coupling: The work definitively proves that Herzberg-Teller vibronic coupling is not a minor correction but a primary driver of ISC/rISC rates in organic emitters, necessitating its inclusion in predictive simulations.
• Open Science: The release of KinLuv allows the community to perform similar multistate kinetic analyses, accelerating the in silico design of high-performance organic light-emitting diodes (OLEDs).

In summary, this paper bridges the gap between qualitative mechanistic proposals and quantitative predictive modeling, establishing that while simplified models work for slow-kinetics TADF emitters, a rigorous inclusion of higher-lying states is mandatory for systems with fast ISC or rIC processes to achieve quantitative agreement with experiment.