Design rules and experimental validation of carbene–metal–amide luminophores: systematic modification of the amide ligand

Abstract

Carbene–metal–amides (CMAs) have emerged as promising thermally activated delayed fluorescence (TADF) emitters for organic light-emitting diodes (OLEDs). Here, we present a comprehensive computational investigation of amide ligand effects on CMA photophysics, examining ca. 70 complexes through density functional theory and its time-dependent extension. Our systematic analysis reveals how structural modifications influence key parameters, including HOMO–LUMO overlap, singlet–triplet energy gaps, oscillator strengths, and metal–ligand bond energies. We demonstrate emission tunability across the visible spectrum through strategic modification of carbazole, indole, carboline, and guanidine-based amides. The computational screening identified promising candidates balancing TADF efficiency with molecular stability. Guided by these predictions, we synthesized two rationally designed complexes with contrasting excited state alignments, and tested their photo- and radioluminescence performance. This combined theoretical–experiment approach establishes clear structure–property relationships for CMA design and demonstrates the effectiveness of computational screening in accelerating OLED material development.

---

Introduction

Organic light-emitting diodes (OLEDs) have emerged as key components in advanced display technologies and energy-efficient lighting solutions.¹ Among promising OLED emitters, carbene–metal–amides (CMAs)^2–21 offer unique advantages through their tunable emission properties and high efficiency. These complexes feature a two-coordinate coinage metal (Au, Cu, Ag), bridging an electron-accepting carbene and electron donating amide ligand.² CMAs achieve high efficiency through thermally activated delayed fluorescence (TADF), which enables harvesting of both singlet and triplet excitons for light emission. In this mechanism, singlet excitons undergo either prompt fluorescence or intersystem crossing (ISC) to form triplet states. These triplet excitons can then absorb thermal energy to undergo reverse ISC (RISC) back to the emissive singlet state, producing delayed fluorescence. The efficiency of this TADF process depends critically on the energy difference between the lowest excited singlet (S₁) and triplet (T₁) states (ΔE_ST).^22,23

Since the introduction of the archetypical CMA1 in 2017,³ extensive studies have been reported, exploring structural modifications to enhance photophysical properties and OLED performance. On the carbene side, variations in cyclic (alkyl)(amino)carbene (CAAC) structure, including ring size alterations and substituent modifications, have demonstrated significant effects on emission wavelengths and quantum yields. Changes to the carbene π-system through benzannulation have enabled access to deep red emission, while modified ring sizes have influenced emission lifetimes and gaps between the highest occupied molecular (HOMO) and lowest unoccupied molecular (LUMO) orbitals.^4–6 On the amide size, systematic modification of the carbazole core through electron-donating and -withdrawing substituents have achieved emission tuning from yellow to deep blue, with some derivatives showing near-unity quantum yields and sub-microsecond delayed fluorescence lifetimes.^4,7

Although these experimental studies have demonstrated the versatility of CMA complexes as OLED emitters, a comprehensive understanding of the structure–property relationships governing their photophysical behavior remains elusive. Building on our previous investigation of CAAC effects in CMA emitters,²⁴ we now present a systematic computational study examining how the amide structural modifications influence key photophysical parameters.

Using density functional theory (DFT)^25,26 and time-dependent DFT (TD-DFT),²⁷ we analyze diverse amide-structural motifs including modified carbazoles, expanded ring systems, and heterocyclic derivatives. Our investigation examines key parameters such as frontier orbital energies, singlet–triplet gaps, oscillator strengths, and bond dissociation energies. The computational analysis guides the synthesis of two rationally designed CMA complexes, demonstrating bright emission in agreement with theoretical predictions. By establishing clear structure–property relationships validated through targeted synthesis, this work provides design principles for high-performance CMA-based OLED materials and demonstrates the effectiveness of computational screening in accelerating emitter development.

Computational details

All calculations were performed using Gaussian 16 ²⁸ with the global hybrid MN15 functional²⁹ and def2-TZVP basis set,^30,31 except the spin–orbit coupling matrix elements (SOCME) were calculated by Orca 6.0.³² For gold atoms, core electrons were treated using a 60-electron relativistic effective core potential.³³ Ground state geometries were fully optimized without symmetry constraints. Excited states were investigated by TD-DFT,²⁷ within the Tamm-Dancoff approximation,³⁴ which helps overcome the typical TD-DFT underestimation of excited states energies.^35,36 The validity of this approach for CMA systems has been previously demonstrated,^8–11 with TD-DFT and unrestricted DFT methods showing only 0.004 eV difference T₁ energies for CMA1.³⁷ Vertical excitations and oscillator strengths were calculated for all complexes, with excited state geometry optimizations and SOC calculations performed for selected cases of particular interest. Electronic structure analysis included Mulliken population analysis to determine metal contributions to frontier orbitals. HOMO–LUMO overlap integral S evaluates the spatial overlap between frontier molecular orbitals and is calculated as:
where Ψ_HOMO and Ψ_LUMO are the normalized HOMO and LUMO wave functions. This integral was computed using Multiwfn program.³⁸

Results and discussion

E-2-1 – The reference complex

Building upon our computational framework, we first analyze the reference complex, e-2-1, an ethyl-substituted CAAC (^EtCAAC) gold carbazole (Fig. 1), which serves as benchmark for investigating amide modifications. This complex has demonstrated superior performance compared to the original CMA1,³ exhibiting 99% photoluminescence (PL) quantum yield (PLQY) in toluene, microsecond-scale excited state lifetime, and excellent stability.²⁴

Fig. 1 Computed structures and properties of e-2-1 complex: (a) optimized ground state S₀ geometry; (b) dipole moments in S₀ and (c) S₁@S₀ state; (d) the HOMO and (e) LUMO orbitals; (f) optimized S₁; (g) optimized T₁; (h) vertical excitation energies indicated by the upward arrows; (i) fluorescence (left) and phosphorescence (right) energies indicated by the downward arrows at optimized S₁ and T₁ geometries.²⁴

The ground state electronic structure reveals characteristic donor–acceptor behavior. The HOMO is predominantly localized on the carbazole amide with minimal (2.8%) gold contribution, while the LUMO distributes across the Au-CAAC moiety with 10.8% metal character (Fig. 1d and e). This orbital arrangement yields a HOMO–LUMO overlap integral of 0.35, favorable for TADF.³⁹ The complex exhibits a substantial ground state dipole moment (10.7 D) oriented along the C–Au–N axis toward the carbene ligand. Notably, upon vertical excitation to S₁@S₀, the dipole moment not only decreases in magnitude (to 3.0 D) but also reverses direction (Fig. 1b and c). This dramatic reversal change reflects the charge-transfer (CT) nature of the excitation, as electron density shifts from the amide to the carbene moiety and may significantly influence the interaction between the complex and its environment in OLED devices.⁴⁰ Vertical excitations analysis shows that both S₀–S₁ and S₀–T₁ transitions are dominated by HOMO-to-LUMO character (>94%), confirming their CT nature. The calculated energies (S₁: 3.17 eV, T₁: 2.88 eV) yield a ΔE_ST of 0.29 eV. Notably, the lowest triplet excitation (³LE) localized to the amide ligand (³LE(A)) corresponds to the second lowest excited triplet (T₂) state at 3.35 eV, positioned above the CT states as required for efficient TADF.⁴¹

Geometry optimization of the excited states reveals significant structural and energetic changes from the ground state. While S₀ and T₁ geometries maintain a near-coplanarity between amide and carbene ligands, the S₁ state prefers a perpendicular arrangement that lies 0.17 eV lower than the coplanar configuration. This structural reorganization has profound implications for the photophysical properties. The ΔE_ST gap decreases from 0.29 eV in the vertical excitation to 0.11 eV when determined from optimized geometries, potentially enhancing RISC efficiency.⁴² Simultaneously, the S₁–S₀ oscillator strength undergoes a dramatic reduction, falling from the ground state S₀–S₁ oscillator strength of 0.1952 to zero in the fully relaxed S₁ geometry. The resulting emission energies are 2.25 eV for fluorescence (S₁–S₀@S₁) and 2.34 eV for phosphorescence (T₁–S₀@T₁), reflecting significant Stokes shifts⁴³ (Fig. 1h and i). The vanishing oscillator strength in the relaxed S₁ geometry, when considered alongside the experimentally observed high PLQY, suggests that emission likely occurs from a higher-energy near-coplanar configuration. The impressive TADF performance of the complex can thus be attributed to its ability to dynamically access geometries that strike a crucial balance between efficient RISC, enabled by the reduced ΔE_ST in the perpendicular configuration, and strong radiative transitions from near-coplanar geometries.

Analysis of the structural stability of the complex through bond dissociation energy calculations reveals robust metal–ligand interactions. The Au–C bond exhibits higher strength (403.2 kJ mol⁻¹) compared to the Au–N bond (375.5 kJ mol⁻¹), with the difference (ΔE = 27.7 kJ mol⁻¹) identifying the metal–amide interface as the more vulnerable point for potential degradation. These values indicate overall strong metal–ligand coordination, crucial for maintaining structural integrity under typical OLED operating conditions.⁴⁴

Dataset selection

Building on the promising properties of e-2-1, we conducted a systematic investigation of amide modifications in ^EtCAAC–Au–amide complexes to establish structure–property relationships. Our design strategy encompassed both validated experimental structures and carefully selected hypothetical derivatives, organized into four main categories (Fig. 2): carbazole substitution patterns, ring-expanded systems, heterocyclic variants, and guanidine-based amides.

Fig. 2 Structural variations of ^EtCAAC–Au–amide complexes investigated in this study, categorized by amide modification type.

1. Carbazole modifications: we systematically functionalized the carbazole's six-membered rings at α, β, γ, and δ positions with diverse substituents including: electron-donating groups (OH, OCH₃), electron-withdrawing groups (CF₃, CN), alkyl groups (CH₃, ^tBu), and halogens (Cl, Br). This comprehensive substitution strategy was motivated by experimental findings showing that ^tBu- and CF₃-substituted carbazoles yield CMA emitters with tunable emission from yellow to deep blue.⁷

2. Ring expansions: we investigated the effect of expanding the carbazole's five-membered ring using heteroatoms (N, O, S) and alkyl groups. Previous work has demonstrated that such modifications enable broad spectral tuning, with PLQYs reaching 89% and successful implementation in yellow, sky-blue, and warm-white OLEDs.⁸

3. Indole and carboline derivatives: following reports of efficient blue emission from azacarbazole-based CMAs,⁹ we explored structurally related heterocycles. This series includes indole and carboline cores with established synthetic accessibility (e.g. skatole,⁴⁵ indoline,⁴⁶ harmane,⁴⁷ pinoline⁴⁸), allowing for potential experimental validation.

4. Guanidine amides: recent demonstrations of exceptional performance of rigid benzoguanidine-based CMAs¹⁰ prompted investigation of related structures. We examined both simple guanidine cores (G1–G6) and extended π-systems (G7–G13), targeting the promising deep-blue emission (424–466 nm), unity PLQY, and sub-microsecond delayed fluorescence demonstrated experimentally.¹⁰

Our computational analysis encompassed 70 complexes, providing systematic insights into how amide structural variations influence key PL parameters including excited state energies, HOMO–LUMO characteristics, and charge-transfer properties. This comprehensive investigation establishes design principles for optimizing CMA emitters, as detailed in subsequent sections.

Structural and electronic analysis

We begin our analysis by examining three fundamental electronic parameters: metal orbital contributions, HOMO–LUMO overlap, metal–ligand bond strengths. These ground state properties provide crucial insights into TADF efficiency, emission characteristics, and device stability.

The gold atom's participation in frontier molecular orbitals varies significantly across the investigated complexes. For the HOMO, primarily localized on the amide ligand, shows Au contributions ranging from 0.0% to 5.4%. The LUMO, distributed over the Au–^EtCAAC moiety, typically exhibits larger Au character (10.0%–15.6%). The reference complex e-2-1 shows moderate Au participation of 2.8% and 10.8% in HOMO to LUMO, respectively.

Systematic modification of the carbazole structure reveals distinct trends in metal orbital contributions. Substitution at the α position generally leads to higher Au participation in frontier orbitals, with complex α_^tBu showing the highest LUMO contribution (15.6%) across all studied compounds. The effect of β-substitution depends strongly on electronic character: electron-donating groups at this position (as in β_OH and β_OCH₃) completely eliminate Au contribution to HOMO (0.0%), while electron-withdrawing substituents lead to moderate Au participation (2.8–3.3%). Substitution at γ- and δ-positions typically results in lower Au contribution to HOMO (γ: 2.4–3.2%, δ: 2.0–2.9%). Ring expansion of the carbazole core maintains moderate HOMO contributions (2.4–4.5%) while showing relatively consistent LUMO participation (11.1%–13.6%). Both indole and carboline derivatives show similar patterns of metal participation, with LUMO contributions (10.2–11.6%) comparable to the reference complex e-2-1. Guanidine-based amides exhibit the widest range of HOMO Au contributions (1.1–5.0%) while maintaining typical LUMO participation (10.3–12.2%). Notably, across all modifications, the variation in metal contribution to HOMO versus the consistency in LUMO reflects the fundamental electronic structure of CMAs: while the HOMO is predominantly localized on the modified amide ligand, the LUMO distributes across the unchanged Au–^EtCAAC moiety.

The HOMO–LUMO overlap integral, crucial for TADF efficiency,⁴⁹ varies from 0.15 to 0.45 across our dataset. The reference complex e-2-1 exhibits a moderate overlap integral of 0.35. Substitution effects are most pronounced at the β position, where electron-donating groups significantly reduce orbital overlap: complexes β_OH and β_OCH₃ show the lowest values in our dataset (0.15 and 0.17, respectively). Ring-expanded complexes maintain overlap integrals (0.34–0.39) similar to the reference compound. Among heterocyclic variants, complex I4 exhibits the highest overlap integral (0.45) in our dataset, while most carboline derivatives show values comparable to e-2-1. Guanidine-based amides (G1–G13) show moderate to low overlap integrals (0.28–0.37). As amide modifications primarily affect the HOMO while leaving the Au-^EtCAAC-centered LUMO largely unchanged, the observed variations in HOMO–LUMO overlap reflect the extent to which different amide structures alter the spatial and energetic characteristics of the HOMO.

To assess potential device stability, we calculated Au–N(amide) and Au–C(carbene) bond dissociation energies. The reference complex e-2-1 exhibits Au–N and Au–C bond energies of 375.5 and 403.2 kJ mol⁻¹, respectively. Across the series, the Au–N bond shows greater variability (274.7–411.7 kJ mol⁻¹) compared to the Au–C bond (373.7–424.6 kJ mol⁻¹), reflecting its greater sensitivity to amide modifications. Electron-withdrawing substituents generally strengthen both metal–ligand bonds, exemplified by complex β_CF₃ with the highest Au–N bond energy of (411.7 kJ mol⁻¹) among carbazole derivatives. This strengthening can be attributed to the reduction of electron density on the amide nitrogen, enhancing the electrostatic interaction with the gold center. Conversely, electron-donating groups tend to weaken this interaction by increasing electron density on the nitrogen. Ring expansions significantly weaken the Au–N bond (274.7–365.2 kJ mol⁻¹) while maintaining relatively strong Au–C bonds (377.0–416.2 kJ mol⁻¹). This weakening is likely due to electronic effects rather than steric factors, with different heteroatoms or groups in the expanded ring altering the electronic structure of the entire π-system. Indole derivatives (I1–I4) show moderate Au–N bond strengths (318.4–378.0 kJ mol⁻¹), while carboline-based complexes (C1–C10) maintain stronger metal–ligand bonds (Au–N: 348.0–403.9 kJ mol⁻¹, Au–C: 395.8–412.2 kJ mol⁻¹). Guanidine-based amides (G1–G13) exhibit lower Au–N bond energies (299.2–375.1 kJ mol⁻¹) compared to the reference complex, which could be explained by the varying degrees of electron delocalization within the guanidine moiety. The markedly different responses of Au–N versus Au–C bonds to amide modifications reflect the localized nature of structural changes: while Au–N interaction is directly influenced by modifications to the amide, the Au–C bond to the unchanged carbene remains relatively stable.

Excitation dynamics

The excited state properties of CMA complexes critically determine their TADF performance and emission characteristics. Our computational analysis reveals that most studied complexes exhibit CT transitions for both S₀–S₁ and S₀–T₁ excitations, dominated by HOMO to LUMO character. This CT character arises from electron migration from the amide-localized HOMO to the Au–carbene centered LUMO. Exceptions appear for cases where electron-withdrawing substituents such as CF₃ are attached on the carbazole ligand, namely the S₀–T₁ transition is characterized by amide-localized (LE(A)), as indicated in Table S1. Higher-lying T₂ states show more diverse behavior: most complexes exhibit LE(A) or mixed CT/LE(A) transitions, while guanidine derivatives with single π-bonds (G1–G5) uniquely exhibit mixed CT/LE(C) character involving the carbene ligand. This variation in T₂ character demonstrates how amide modifications can influence higher excited states while maintaining the fundamental CT nature of S₁ and T₁.

Our computational results reveal systematic tuning of excited state energies through structural modification (Fig. 3). S₁ energies span from 2.38 eV in the nitrogen-expanded carbazole complex (R_N) to 3.91 eV in the α-substituted complex (α_CF₃), corresponding to emission wavelengths from 521 nm to 317 nm before accounting for Stokes shift effects. T₁ states follow a similar pattern, ranging from 2.04 eV (R_N) to 3.32 eV (α_CF₃), with S₁/T₁ energy ratios remaining notably consistent between 1.06–1.22 across the series.

Fig. 3 Vertical excitation energies (eV) of CMA complexes with different amide ligands, including carbazole modifications (yellow), ring expansions (blue), carboline derivatives (green), indole derivatives (orange), and guanidine amides (purple).

Another excited state that warrants careful consideration relative to S₁ and T₁ states is T₂. For efficient TADF, the optimal energetic hierarchy is T₂ > S₁ > T₁. This ordering ensures that thermal equilibration occurs primarily between CT states S₁ and T₁, without interference from higher triplet manifolds. If T₂ is positioned energetically below S₁, it may serve as a competing decay channel, thereby disrupting the desired S₁ ↔ T₁ equilibrium essential for TADF. In such cases, excitons in the S₁ state can undergo competing ISC to the lower-lying T₂ state, removing population from the TADF-active S₁ state. Additionally, thermal activation from T₁ to S₁ becomes less favorable when T₂ provides an alternative, lower-energy triplet state. Furthermore, T₂ states often exhibits distinct orbital character, potentially introducing additional non-radiative decay pathway. In the systems under investigation, T₂ state energies (2.79–3.94 eV) typically lie above S₁, thereby supporting favorable conditions for TADF. However, several compounds show T₂ states below their respective S₁ energies. For instance, in γ_OH, T₂ (2.89 eV) lies below S₁ (2.95 eV), and similar inversion occurs with many other substitution patterns. This unfavorable energy ordering could compromise TADF efficiency. The occurrence of such cases, particularly with electron-withdrawing substituents, highlights a key molecular design challenge: modifications targeting blue emission through raised S₁ energies must simultaneously maintain appropriate excited state ordering.

Analysis of other structural modifications reveals distinct trends in excited state ordering. Ring-expanded complexes maintain favorable T₂ > S₁ ordering, with R_N and R_N(CH₃) exhibiting the lowest S₁ energies in our dataset (2.38–2.39 eV), though the series extends to higher energies up to 3.31 eV in R_SO₂. Indole derivatives (I1–I4) similarly preserve T₂ above S₁ with substantial gaps (T₂–S₁ ≥ 0.19 eV), though at typically higher S₁ energies (2.87–3.25 eV). Carboline derivatives, despite their promise as blue emitters with S₁ energies up to 3.48 eV, show several cases of unfavorable energy ordering, particularly in C7 and C8, where T₂ lies below S₁ by 0.10–0.12 eV. In contrast, guanidine-based complexes (G1–G13) consistently maintain T₂ above S₁ across a wide range of S₁ energies (2.64–3.56 eV), suggesting this modification strategy may be particularly robust for maintaining appropriate state ordering while tuning emission wavelength.

The TADF-critical ΔE_ST gap and S₀–S₁ oscillator strength show systematic variation with structural modification (Fig. 4). Among carbazole derivatives, α-substitution produces the widest range of ΔE_ST values, from α_^tBu (0.17 eV) to α_CF₃ (0.59 eV). Ring expansions generally maintain moderate ΔE_ST gaps (0.26–0.35 eV), while guanidine derivatives, particularly G9–11 and G13, achieve consistently small gaps (0.20–0.21 eV). These variations correlate with HOMO–LUMO overlap integrals: complexes with smaller overlaps typically show reduced ΔE_ST gaps but often at the cost of diminished oscillator strength. Notably, β_OH and β_OCH₃ show minimal overlaps (0.15–0.17) and moderate ΔE_ST (0.29–0.30 eV), with reasonably high oscillator strengths (≈0.19), but their nearly degenerate S₁ and T₂ states (ΔE ≈ 0.03 eV) may complicate TADF performance. Complex γ_^tBu maintains moderate overlap integral (0.35) and achieves high oscillator strength (≈0.22) while keeping ΔE_ST manageable (0.29 eV), with favorable T₂–S₁ gap of 0.13 eV. However, despite promising ΔE_ST (0.31 eV) and high oscillator strength (≈0.23), carboline derivative C5 shows unfavorable state ordering with T₂ (3.22 eV) below S₁ (3.29 eV). These results highlight the multiple requirements for efficient TADF emitter design: optimal complexes must balance moderate ΔE_ST gaps for efficient RISC, maintain sufficient oscillator strength for strong radiative decay, and crucially, preserve favorable excite state ordering with T₂ above S₁.

Fig. 4 Comparison between overlap integral, ΔE_ST, and oscillator strength of CMA complexes with different amide ligands, including carbazole modifications (yellow), ring expansions (blue), carboline derivatives (green), indole derivatives (orange), and guanidine amides (purple).

Electrostatic dipole moments provide insight into potential reorganization energies and TADF efficiency.⁴⁰ Ground state dipole moments in all complexes align along the C(carbene)–Au–N(amide) axis towards the carbene, mirroring e-2-1 (10.7 D), but with magnitudes varying widely (4.1–17.0 D). Notably, guanidine-based complexes (G1–G6) show the smallest S₀ dipole moments (4.1–5.4 D), while some γ- and δ-substituted carbazoles exhibit the largest (e.g. δ_CF₃ with 17.0 D; γ_Cl, γ_Br, and γ_CF₃ with 14.2–16.0 D). Upon S₀–S₁ excitation, most complexes exhibit a reversal in dipole moment direction, consistent with CT character, and often a significant change in magnitude. The S₁@S₀ dipole moments range from −8.0 D to 4.9 D. Large changes are observed in complexes like (α_CF₃) (β_OCH₃) (from 8.0 D to −7.2 D) and γ_Br (from 14.5 D to −0.8 D). These substantial reorganizations of charge distribution may impact TADF performance through enhanced non-radiative decay or altered solid-state packing interactions.⁴⁰

Structure–property relationships

Our comprehensive study reveals intricate relationships between structural modifications and key photophysical properties of CMA complexes. We analyze these relationships across various structural motifs. Table 1 and S2 illustrate comparative analysis providing valuable insights into distinct trends in how different amide structures affect calculated parameters.

Table 1 Comparative analysis of the general effect of the amide structure on calculated parameters related to PL characteristic

↗ increasing, ↘ decreasing, ∼ maintaining,/or. Red color: electron-donating group. Blue color: electron-withdrawing group. Green color: guanidine amide with one π bond (G1–G6). Purple color: guanidine amide with increasing conjugation (G7–G13).

---

Carbazole α-substitution offers the widest tunability but with important trade-offs for TADF design. Electron-donating groups (OH, OCH₃, ^tBu) maintain favorable HOMO–LUMO overlap and reduce ΔE_ST gaps, supporting efficient RISC, while enhancing oscillator strength for bright emission. However, electron-withdrawing groups (CF₃, Cl, Br) achieve substantial blue-shifts (up to 3.91 eV) but at the cost of increased ΔE_ST gaps (up to 0.59 eV) and reduced oscillator strength. The α-position shows the strongest electronic coupling to the metal center, with α_^tBu exhibiting the highest Au contribution to frontier orbitals (15.6% LUMO). For blue emission applications, α-substitution requires careful balance between achieving target energies and maintaining TADF efficiency.

β-substitution provides unique electronic control through dramatic HOMO–LUMO decoupling. Electron-donating groups (β_OH, β_OCH₃) create the most spatially separated donor–acceptor systems in our dataset (overlap integrals 0.15–0.17), potentially enabling very small ΔE_ST gaps while maintaining reasonable oscillator strengths (ca. 0.19). However, their nearly degenerate S₁ and T₂ states may complicate TADF dynamics. Electron-withdrawing groups offer substantial blue-shifts (up to 3.59 eV with β_CF₃) while preserving electronic coupling, making this position ideal for high-energy blue emitters. Uniquely among all positions, β-substitution universally strengthens metal–ligand bonds, suggesting enhanced operational stability for device applications.

γ-substitution emerges as the “sweet spot” for TADF optimization, offering emission tuning without compromising key electronic parameters. Unlike other positions, γ-substitution maintains consistent HOMO–LUMO overlap (0.34–0.36) and favorable ΔE_ST gaps (0.27–0.31 eV) across all substituents, while uniquely preserving or enhancing oscillator strength. Complex γ_^tBu exemplifies this balance with high oscillator strength (0.22), moderate ΔE_ST (0.29 eV), and favorable T₂–S₁ ordering. The position enables systematic emission tuning (2.95–3.55 eV) while maintaining TADF-critical parameters, making it ideal for fine-tuning emission color without sacrificing efficiency.

δ-substitution offers similar benefits to γ-substitution but with enhanced ΔE_ST control. Electron-donating groups achieve the smallest ΔE_ST gaps in the carbazole series (δ_OH: 0.24 eV), making this position attractive for maximizing RISC efficiency. However, electron-withdrawing groups produce detrimental large gaps (δ_CF₃: 0.42 eV) with reduced oscillator strength, limiting blue emission applications. The position provides a viable alternative to γ-substitution when minimal ΔE_ST is prioritized over oscillator strength, though the latter trade-off must be carefully considered for overall TADF performance.

Ring expansion of the five-membered carbazole ring provides the most dramatic red-shifts in our dataset but with significant stability trade-offs. Nitrogen incorporation (R_N, R_N(CH₃)) achieves the lowest S₁ energies (2.38–2.39 eV) for accessing orange-red emission, while maintaining favorable ΔE_ST gaps for TADF. However, these systems suffer from substantially weakened Au–N bonds (274.7–277.1 kJ mol⁻¹ vs. 375.5 kJ mol⁻¹ reference) and reduced oscillator strength, potentially compromising device stability and brightness. The R_SO₂ system offers a compromise, approaching reference S₁ energy while providing the smallest ΔE_ST gap (0.26 eV) in the series. Ring expansion is most suitable for red-shifted emission applications where the stability trade-offs are acceptable.

Indole derivatives split into two distinct categories with contrasting design implications. Complexes I1–I3 maintain TADF-favorable properties similar to the reference (ΔE_ST: 0.27–0.29 eV, S₁: 3.15–3.25 eV) but with reduced oscillator strength, making them suitable for blue emission with moderate brightness. In stark contrast, I4 represents an outlier with the highest HOMO–LUMO overlap in our dataset (0.45) and enhanced oscillator strength (0.25), but suffers from a prohibitively large ΔE_ST gap (0.52 eV) that likely prevents efficient TADF. The indole scaffold offers a reliable platform for blue emitters (I1–I3), while I4 serves as a cautionary example of how structural modifications can disrupt TADF balance.

Carboline derivatives demonstrate how nitrogen position critically affects TADF performance, creating two distinct design pathways. The C5–C6 series achieves excellent oscillator strength (0.20–0.23) with moderate ΔE_ST gaps, making them promising blue emitters despite slightly unfavorable T₂ ordering. Conversely, C7–C8 suffer from poor oscillator strength (0.08–0.09) and large ΔE_ST gaps (0.44–0.45 eV), effectively eliminating TADF capability. The carboline scaffold offers high-performance blue emission potential (C5, C6) but requires careful nitrogen positioning to avoid TADF-detrimental configurations. Enhanced metal–ligand bonding across the series suggests improved stability compared to other heterocyclic modifications.

Guanidine amides reveal how π-conjugation extent fundamentally determines emitter performance, creating two complementary design strategies. Single π-bond systems (G1–G6) target deep-blue emission (3.33–3.56 eV) but suffer from large ΔE_ST gaps (up to 0.41 eV) and poor oscillator strength, limiting TADF efficiency. Extended conjugation systems (G7–G13) unlock exceptional TADF performance with small ΔE_ST gaps (0.20–0.21 eV) and tunable emission (2.64–3.27 eV), with G13 demonstrating optimal balance (ΔE_ST: 0.21 eV, oscillator strength: 0.18). The guanidine platform offers a unique design lever – π-extension – to transition from stable blue phosphors (G1–G6) to high-performance TADF emitters (G7–G13).

Implications and design strategy

The systematic structure–property analysis reveals a clear hierarchy of design strategies for optimizing CMA emitters. For emission tuning, γ-substitution emerges as the most versatile approach, offering systematic color control (2.95–3.55 eV) without compromising TADF parameters, while ring expansion provides access to deep-red emission when stability trade-offs are acceptable. For maximizing TADF efficiency, β-electron-donating groups and guanidine π-extension (G7–G13) achieve the smallest ΔE_ST gaps, while α-electron-donating groups optimize oscillator strength. For blue emission applications, carboline C5–C6 types and γ-/δ-electron-withdrawing substitutions provide high-energy access with maintained efficiency, avoiding problematic configurations like C7–C8 or I4.

Based on this framework, computational screening identifies five leading candidates that exemplify optimal design principles. Complex γ_^tBu demonstrates the γ-position advantage with balanced parameters across all metrics: moderate ΔE_ST (0.29 eV), high oscillator strength (0.22), and favorable T₂–S₁ ordering. α_OH achieves the highest oscillator strength in our dataset (0.25) while maintaining blue emission capability (3.21 eV). Carboline C6 balances strong emission properties (oscillator strength: 0.20) with enhanced stability, though T₂ positioning requires consideration. Guanidine complexes G12 and G13 represent the success of the π-extension strategy, combining small ΔE_ST gaps (0.21–0.27 eV) with robust T₂ > S₁ ordering essential for uncompromised TADF.

To validate these computational predictions and understand the emission mechanisms, excited state geometry optimizations were performed for all five candidates. The optimized structures (Fig. 5) reveal distinct photophysical behaviors that explain their relative performance. Complexes γ_^tBu, C6, G12, and G13 exhibit the expected TADF-favorable rotation from coplanar S₀/T₁ geometries to perpendicular S₁ configuration, reducing ΔE_ST to 0.11–0.20 eV (Table 2). This rotation mechanism enables thermal S₁ ↔ T₁ equilibration essential for TADF, though requiring geometric reorganization (0.06–0.17 eV) to achieve maximum oscillator strength for emission. In contrast, α_OH demonstrates why α-substitution requires careful design consideration: steric constraints from the proximal hydroxyl group prevent S₁ rotation, maintaining coplanar geometry and resulting in a large ΔE_ST (0.26 eV) that impedes RISC. All complexes maintain beneficial ³LE positioning (3.13–3.37 eV) above CT states, confirming appropriate excited state ordering for TADF applications.

Fig. 5 S₀, S₁, and T₁ optimized geometries.

Table 2 S₁ and T₁ excited state optimizations and SOCME

		α_OH	γ_^tBu	C6	G12	G13
a Vertical S1 and T1 excitations.b Optimized S1 and T1 geometries.c Optimized T1 geometry.d Carbene and amide fixed coplanar.
Vertical S1 excitation energy (eV)		3.21	3.09	3.31	3.16	3.17
Vertical T1 excitation energy (eV)		2.91	2.80	3.02	2.89	2.96
ΔEST ^a (eV)		0.30	0.29	0.29	0.27	0.21
Energy relative to optimized S0 (eV)	Optimized coplanar S1	2.84	2.81	3.02	2.79	2.84
	Optimized rotated S1	2.87	2.64	2.95	2.66	2.78
	Optimized coplanar T1	2.58	2.53	2.75	2.52	2.66
ΔEST ^b (eV)		0.26	0.11	0.20	0.14	0.12
^3LE energy^c (eV)		3.15	3.13	3.14	3.28	3.37
Maximum S1–S0 oscillator strength^d		0.0820	0.1670	0.1518	0.1136	0.1322
Fluorescence (S1–S0@S1) (eV)		2.13	2.16	2.32	2.05	2.31
Phosphorescence (T1–S0@T1) (eV)		2.18	2.24	2.45	2.08	2.31
^1CT and ^3CT SOCME (cm^−1)		100.79	3.77	4.03	6.20	3.65

---

Spin–orbit coupling analysis provides additional mechanistic insight into these performance differences. The calculated SOCMEs between ¹CT and ³CT states reveal significant variations across the series (3.65–100.79 cm⁻¹, Tables 2 and S3). Complex α_OH exhibits the highest SOCME (100.79 cm⁻¹), indicating rapid ISC from singlet to triplet state. However, its large ΔE_ST (0.26 eV) and inability to adopt the perpendicular S₁ geometry hinder efficient RISC, negatively impacting TADF performance. In contrast, γ_^tBu, C6, G12, and G13 show significantly smaller SOCMEs (3.65–6.20 cm⁻¹) combined with favorable small ΔE_ST gaps (0.11–0.20 eV) that enable efficient RISC. Notably, G13 achieves optimal balance with the smallest SOCME (3.65 cm⁻¹) and ΔE_ST gaps (0.12 eV), consistent with its superior experimental TADF performance (vide infra). These results suggest that moderate SOC strength is preferable for CMA emitters, providing sufficient ISC/RISC rates while minimizing competing non-radiative decay pathways.

Synthesis and testing of promising candidates

Following our computational analysis identifying C6 and G13 as promising candidates, the closely related complexes C5 and G13 (Fig. 6) were synthesized and fully characterized to validate our theoretical predictions. Complex C5 represents an excellent structural analogue of C6 with nearly identical calculated properties, providing effective experimental validation of our computational framework. Both complexes are obtained from the (^Et2CAAC)AuCl and benzoguanidine (G13) or harmane (C5) in the presence of the KO^tBu base in high yields. All complexes are off-white solids having high stability in air for months and good solubility in polar organic solvent while only sparingly soluble in hexane. Both C5 and G13 have been characterised by ¹H and ¹³C NMR (Fig. S1–S4) and high-resolution mass spectroscopy. The thermal stability of the complexes was evaluated with thermogravimetric analysis (TGA, under nitrogen). The decomposition temperature (T_d) for gold complex C5 is 273 °C and 303 °C for G13 which is close similar to 304 °C for analogous carbazolide complex e-2-1.

Fig. 6 Single crystal X-ray structure for complex G13 (left) and two independent molecules A (middle) and B (right) for complex C5. The ellipsoids are shown at 50% probability while H atoms are omitted for clarity.

The molecular structure for CMA complexes C5 and G13 were confirmed by single crystal X-ray diffraction (Fig. 6). Both C5 and G13 have a similar linear geometry for the gold atom with negligible deviation up to 2° from the ideal 180° angle for linear geometry. Unlike complex e-2-1 ²⁴ having a C_2v symmetry for carbazolide amide donor moiety, complexes C5 and G13 possess unsymmetrically substituted amide donor. This results in crystallization of the complex C5 with two independent molecules A and B in the unit cell which are different by conformation of the ethyl-group for the ^EtCAAC–carbene π-acceptor ligand and relative orientation with the amide π-donor ligand. For instance, molecule B possesses methyl C-atom C25 pointed towards the center of the 2,6-diisopropylaniline moiety of carbene whereas C25 positioned in the space between of two ethyl groups of the carbene for molecule A. Complex G13 has only one independent molecules in the unit cell. Both Au–C and Au–N bond lengths for C5 and G13 are up to 0.02 Å longer compared to benchmark complex e-2-1, indicating greater separation between donor and acceptor ligands for the new CMA materials and influencing their photophysical behaviour, vide infra. The torsion angle (C1–N1–N2–C23, Fig. 6) between carbene and carbazole ligands varies in the range from 3.0(1) to 10.7(1)° which is close for complex e-2-1 (ca. 15.1(1)°), indicating near co-planar orientation of the ligands. Analysis of the intermolecular interactions between neighbouring molecules of complexes C5 and G13 reveals only weak C–H⋯π and C–H⋯N contacts resulting in formation of the 3D-network.

The redox behaviour of complexes C5 and G13 was analysed in THF solution with profiles shown in Fig. S6 and data collected in Table S5. Both complexes exhibit a quasi-reversible, carbene ligand-centred reduction process at E_1/2 value at −2.80 V. All oxidation processes are centered on the amide and irreversible with E_p values at ca. +0.46 V (Table S5). This results in close similar values for LUMO at ca. 2.70 eV for both complexes but slightly more stabilised HOMO energy level at −5.66 eV for C5 compared to −5.56 eV for G13. The UV-vis absorption spectra in solvents with different polarity (Fig. S5, Table S4) show two major absorption bands: high-energy (ca. 300 nm) and low-energy absorption (ca. 360 nm) band. Complex C5 exhibits only well-resolved and narrow absorption bands for both high- and low-energy absorption bands with no solvatochromism effect. We ascribed both bands to π–π* transition localised on the amide-donor ligand, indicating absence of the low energy CT band. Unlike C5, complex G13 possesses a broad L(M)L CT low-energy absorption band, which experience 25 nm blue-shift from methylcyclohexane (MCH) to dichloromethane (DCM, Fig. S5, Table S4). Such a negative solvatochromism for complex G13 is characteristic for the CMA materials having bright TADF emission mechanism, vide infra.

The PL spectra were measured in neat crystals and polystyrene (PS) matrix at 295 K and 77 K (Table 3 and Fig. 7). Both complexes C5 and G13 emit blue light at 451 and 460 nm, respectively, which is up to 31 nm blue-shift compared with the e-2-1 in PS matrix²⁴ (484 nm, Table 3). The PL profile for harmane complex C5 possesses poorly resolved vibronic structure (Fig. 7a top) which is becoming apparent upon cooling to 77 K (Fig. 7a bottom). Unlike C5, PS films of benzoguanidinato complex G13 possess a broad CT profile at all temperatures from 295 K to 16 K (Fig. 7b) thus demonstrating the PL behaviour similar to complex e-2-1. The CT-state energy for complexes C5 and G13 was estimated at ca. 3.06 eV from the blue onset of the PL profile in PS matrix at 295 K (Table 3). The excited state lifetime of benzoguanidinato CMA complex G13 is 0.38 μs which is three orders of magnitude shorter compared to harmane complex C5 (up to 298 μs) and nearly two-time shorter compared to carbazolide CMA complex e-2-1 (ca. 1.0 μs). Complex G13 possesses a radiative rate of 2.6 × 10⁶ s⁻¹ which is two-orders of magnitude faster compared to C5 (2 × 10⁴ s⁻¹, Table 3) thanks to a short excited state lifetime and near unity PLQY for G13 and only 40% PLQY for C5. Such drastic differences in PL performance clearly indicate a different luminescence mechanism governing the PL behaviour of the complexes C5 and G13.

Fig. 7 PL (em – emission; ex – excitation spectrum) of C5 (a) and G13 (b) in PS matrix with 1% concentration by weight at 295 K (top), and 77 K or 16 K (bottom). The emission profiles in MeTHF frozen glass at 77 K, revealing a contribution of the ³LE(A) phosphorescence (c); excited state lifetime vs. temperature plot for G13 in PS matrix (d) and Arrhenius plots of ln(k_TADF) vs. 1/T and fitted according to equation: .

Table 3 Photophysical properties of e-2-1,²⁴ C5 and G13 complexes in various media

	λem (nm)	τ (μs)	Φ^a (%)	kr ^b(10^5 s^−1)	knr ^c(10^5 s^−1)	CT/^3LE(A)/ΔEST ^d (eV)	λem (nm, 77 K)	τ (μs, 77 K)
a Quantum yields determined using an integrating sphere.b Radiative rate constant kr = Φ/τ.c Nonradiative constant knr = (1 − Φ)/τ.d CT/^3LE(A) energies based on the onset values of the emission spectra blue edge in MeTHF glasses at 77 K and 295 K.
MeTHF frozen glass
C5	—	—	—	—	—	—/2.87	450	2522
G13	—	—	—	—	—	—/3.33	402	67 (90%), 1100 (10%)
Polystyrene Matrix (1 wt%)
e-2-1	484	1.07	82	7.66	1.68	2.92/2.95/−0.03	458	48.0 (63%), 188.3 (32%)
C5	451	40 (30%), 298 (70%)	40	0.02	0.03	3.06/2.87/+0.19	449	401 (44%), 1162 (32%)
G13	460	0.38	99	26	0.26	3.07/3.33/-0.26	448	55

---

To reveal the energy of the ³LE state, we performed the PL spectra for both complexes in 10⁵M MeTHF frozen glass at 77 K (Fig. 7c). Both complexes possess a vibronically resolved PL profile indicative of a dominant phosphorescence from the ³LE(A) at 2.87 eV for C5 and 3.33 eV for G13. Both complexes possess the excited state lifetime in millisecond range supporting assignment of the phosphorescence at 77 K and in line with our previous observations for CMA complexes emitting phosphorescence from the locally excited triplet ³LE(A).¹⁰ The energy difference between CT and ³LE(A) states, ΔE_ST, is larger for compound G13 (−0.26 eV) compared with C5 (+0.19 eV), while ³LE(A) state is higher in energy compared with CT state for G13 (Table 3). Such experimental observation unequivocally corroborates our theoretical calculations above. For instance, complex G13 has the same excited CT and LE-states orientation with the reference compound e-2-1 (see Fig. 1), where triplet ³LE state is a T₂ state laying above the CT singlet S₁ and triplet T₁. Theoretical calculations predict the energy of the triplet ³LE(A) to be 3.37 eV and energy gap of −0.21 eV (Table 2) which is in excellent agreement with the experimental value of 3.33 eV for ³LE(A) and −0.26 eV for the energy gap in complex G13. Therefore, theory and experiment enabled us to assign TADF emission mechanism for the complex G13. We also performed a variable-temperature lifetime measurement to estimate the activation energy barrier, ΔE_a, for G13 in PS film. On cooling to 16 K, the PL profile remains broad and unstructured (Fig. 7b and d) while the emissive lifetime increases from 0.38 μs at 295 K to 55 μs at 16 K (Table 3, Fig. 7d). The ΔE_a for G13 was estimated as 49.1 meV by an Arrhenius plot analysis of the varied temperature data with ln k_TADF vs. 1/T, supporting high radiative rate of TADF.

Theoretical calculations for complex C5 predicted that both CT and ³LE(A) triplets are below in energy compared to CT singlet S₁ state. This may result in a mixed character emission from CT and ³LE states simultaneously which is in a good agreement with the experimental observation for PS film of C5 – a rather broad CT emission profile with features of vibronic structure even at room temperature (Fig. 7a, Table 3).

Spectacular photophysical characteristics for complex G13 (unity PLQY and short excited state lifetime of 380 ns) spurred our interest to test the applied potential of G13 in radioluminescence experiment. This is important for development of the advanced scintillators for detecting high-energy X-ray radiation (radioluminescence, RL). Crystals of G13 were irradiated in an X-ray cabinet at 350 kV (max) and 11.4 mA current, with emission spectra measured using a fibre-optic cable connected to a fluorimeter. Complex G13 exhibits a RL emission spectra similar to its PL profile (Fig. 8). Radiostability was assessed by monitoring emission intensity at the compound's emission maxima every 10 seconds during constant irradiation at 16.357 Gy min⁻¹ (total dose ca. 980 Gy). Complex G13 shows excellent stability with LT₇₅ = 1 h (where LT₇₅ represents the time required for RL intensity to decrease by 25% under constant 350 kV X-ray radiation). The radioluminescence stability for complex G13 sample is comparable with the industrial standard BGO (bismuth germanium oxide, powder sample) as shown in Fig. 8.

Fig. 8 Overlay of photoluminescence (PL, excited at 360 nm) and radioluminescence emission spectra for G13 (left); emission intensity decay for G13 and BGO (industry standard sample) under constant 350 kV X-ray irradiation (right).

Conclusions

This work presents a comprehensive computational investigation of amide ligand effects on CMA photoluminescence properties, examining approximately 70 complexes to establish structure–property relationships critical for OLED applications. Our systematic analysis reveals how specific structural modifications – from carbazole substitutions to heterocyclic variations – influence key photophysical parameters that determine emission properties and device performance. The study demonstrates that efficient TADF requires precise control of multiple interdependent factors: HOMO–LUMO overlap, S₁–T₁ energy gap, oscillator strength, and metal–ligand bond strengths. Importantly, T₂ state positioning relative to S₁ is crucial for preventing competing decay pathways, while high bond dissociation energies support operational stability.

Our computational screening reveals remarkable emission tunability across the visible spectrum (2.38–3.91 eV), with systematic trends in how structural modifications affect emission energy. Electron-withdrawing substituents on carbazole consistently blue-shift emission, while electron-donating groups show position-dependent effects: ring expansions and extended π-conjugation provide routes to red-shifted emission, while minimal π-systems enable access to blue regions.

The predictive power of our computational approach is validated through the successful synthesis of two rationally designed complexes, C5 and G13, which exhibit bright phosphorescence and TADF emission, respectively, as predicted by their calculated electronic structures. These complexes, strategically selected for their contrasting excited-state alignments, demonstrate how computational insight can guide the targeted development of high-performance emitters. Complex G13 shows superior deep-blue TADF luminescence with radiative rates up to 2.6 × 10⁶ s⁻¹ enabling excellent radioluminescence stability under constant 350 kV X-ray radiation.

This combined theoretical–experimental study establishes a robust framework for CMA emitter design, providing clear guidelines for optimizing TADF performance through structural modification. The validated computational methodology offers an efficient pathway for screening candidate structures before synthesis, accelerating the development of next-generation OLED materials. These insights advance the broader goal of achieving highly efficient, stable, and color-tunable displays and lighting devices based on TADF technology.

Author contributions

Conceptualization, A. S. R. and M. L.; methodology, A. S. R. and M. L.; investigation, N. L. P.; resources, A. S. R. and M. L.; writing—original draft preparation, N. L. P.; synthesis A. K.; crystallography and PL characterization A. K. and A. S. R.; radioluminescence A. B.; writing—review and editing, A. S. R. and M. L.; visualization, N. L. P.; supervision, A. S. R. and M. L. All authors have read and agreed to the published version of the manuscript.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting this article have been included as part of the SI.

Calculated parameters and experimental details, coordinates of optimized structures, cif-files and check CIF for complexes C5 and G13. See DOI: https://doi.org/10.1039/d5qi01245g.

CCDC 2454948 and 2454949 contain the supplementary crystallographic data for this paper.^50a,b

Acknowledgements

DFT computations were made possible by use of the Finnish Grid and Cloud Infrastructure resources (urn:nbn:fi:research-infras-2016072533). N. L. P. acknowledges the Doctoral Programme in Science, Forestry and Technology (Lumeto, University of Eastern Finland), decision 401/2023. A. S. R acknowledges the support from the Royal Society (grant no. URF\R1\180288, RGF\EA\181008, URF\R\231014), EPSRC (grant code EP/K039547/1 and APP46952). M. L. acknowledges the Academy of Finland Flagship Programme, Photonics Research and Innovation (PREIN), decision 320166. We thank Dr Louise Natrajan, EPSRC and University of Manchester for access the Centre for Radiochemistry Research National Nuclear User's Facility (NNUF, EP/T011289/1) to use FLS-1000 fluorometer.

References

1. B. Geffroy, P. Roy and C. Prat, Organic light-emitting diode (OLED) technology: materials, devices and display technologies, Polym. Int., 2006, 55, 572–582.
2. J. Eng, S. Thompson, H. Goodwin, D. Credgington and T. J. Penfold, Competition between the heavy atom effect and vibronic coupling in donor–bridge–acceptor organometallics, Phys. Chem. Chem. Phys., 2020, 22, 4659–4667.
3. D. Di, A. S. Romanov, L. Yang, J. M. Richter, J. P. Rivett, S. Jones, T. H. Thomas, M. A. Jalebi, R. H. Friend, M. Linnolahti, M. Bochmann and D. Credgington, High-performance light-emitting diodes based on carbene-metal-amides, Science, 2017, 356, 159–163.
4. R. Hamze, J. L. Peltier, D. Sylvinson, M. Jung, J. Cardenas, R. Haiges, M. Soleilhavoup, R. Jazzar, P. I. Djurovich, G. Bertrand and M. E. Thompson, Eliminating nonradiative decay in Cu(I) emitters: >99% quantum efficiency and microsecond lifetime, Science, 2019, 363, 601–606.
5. R. Jazzar, M. Soleilhavoup and G. Bertrand, Cyclic (Alkyl)- and (Aryl)-(amino)carbene Coinage Metal Complexes and Their Applications, Chem. Rev., 2020, 120, 4141–4168.
6. F. Chotard, V. Sivchik, M. Linnolahti, M. Bochmann and A. S. Romanov, Mono- versus Bicyclic Carbene Metal Amide Photoemitters: Which Design Leads to the Best Performance?, Chem. Mater., 2020, 32, 6114–6122.
7. P. J. Conaghan, C. S. B. Matthews, F. Chotard, S. T. E. Jones, N. C. Greenham, M. Bochmann, D. Credgington and A. S. Romanov, Highly efficient blue organic light-emitting diodes based on carbene-metal-amides, Nat. Commun., 2020, 11, 1758.
8. A. S. Romanov, S. T. E. Jones, Q. Gu, P. J. Conaghan, B. H. Drummond, J. Feng, F. Chotard, L. Buizza, M. Foley, M. Linnolahti, D. Credgington and M. Bochmann, Carbene metal amide photoemitters: tailoring conformationally flexible amides for full color range emissions including white-emitting OLED, Chem. Sci., 2020, 11, 435–446.
9. A.-P. M. Reponen, F. Chotard, A. Lempelto, V. Shekhovtsev, D. Credgington, M. Bochmann, M. Linnolahti, N. C. Greenham and A. S. Romanov, Donor N-Substitution as Design Principle for Fast and Blue Luminescence in Carbene-Metal-Amides, Adv. Opt. Mater., 2022, 10, 2200312.
10. A. C. Brannan, H.-H. Cho, A.-P. M. Reponen, S. Gorgon, N. L. Phuoc, M. Linnolahti, N. C. Greenham and A. S. Romanov, Deep-Blue and Fast Delayed Fluorescence from Carbene–Metal–Amides for Highly Efficient and Stable Organic Light-Emitting Diodes, Adv. Mater., 2024, 36, 2404357.
11. A. S. Romanov, S. T. E. Jones, L. Yang, P. J. Conaghan, D. Di, M. Linnolahti, D. Credgington and M. Bochmann, Mononuclear Silver Complexes for Efficient Solution and Vacuum-Processed OLEDs, Adv. Opt. Mater., 2018, 6, 1801347.
12. P. J. Conaghan, S. M. Menke, A. S. Romanov, S. T. E. Jones, A. J. Pearson, E. W. Evans, M. Bochmann, N. C. Greenham and D. Credgington, Efficient Vacuum-Processed Light-Emitting Diodes Based on Carbene–Metal–Amides, Adv. Mater., 2018, 30, 1802285.
13. A. S. Romanov, L. Yang, S. T. E. Jones, D. Di, O. J. Morley, B. H. Drummond, A.-P. M. Reponen, M. Linnolahti, D. Credgington and M. Bochmann, Dendritic Carbene Metal Carbazole Complexes as Photoemitters for Fully Solution-Processed OLEDs, Chem. Mater., 2019, 31, 3613–3623.
14. S. Shi, M. C. Jung, C. Coburn, A. Tadle, M. R. D. Sylvinson, P. I. Djurovich, S. R. Forrest and M. E. Thompson, Highly Efficient Photo- and Electroluminescence from Two-Coordinate Cu(I) Complexes Featuring Nonconventional N-Heterocyclic Carbenes, J. Am. Chem. Soc., 2019, 141, 3576–3588.
15. R. Hamze, S. Shi, S. C. Kapper, D. S. M. Ravinson, L. Estergreen, M.-C. Jung, A. C. Tadle, R. Haiges, P. I. Djurovich, J. L. Peltier, R. Jazzar, G. Bertrand, S. E. Bradforth and M. E. Thompson, “Quick-Silver” from a Systematic Study of Highly Luminescent, Two-Coordinate, d¹⁰ Coinage Metal Complexes, J. Am. Chem. Soc., 2019, 141, 8616–8626.
16. M. B. Gildner and T. W. Hudnall, Cyclic (aryl)(amido)carbenes: pushing the π-acidity of amidocarbenes through benzannulation, Chem. Commun., 2019, 55, 12300–12303.
17. M. Gernert, L. Balles-Wolf, F. Kerner, U. Müller, A. Schmiedel, M. Holzapfel, C. M. Marian, J. Pflaum, C. Lambert and A. Steffen, Cyclic (Amino)(aryl)carbenes Enter the Field of Chromophore Ligands: Expanded π System Leads to Unusually Deep Red Emitting CuI Compounds, J. Am. Chem. Soc., 2020, 142, 8897–8909.
18. J. Feng, E. J. Taffet, A.-P. M. Reponen, A. S. Romanov, Y. Olivier, V. Lemaur, L. Yang, M. Linnolahti, M. Bochmann, D. Beljonne and D. Credgington, Carbene-Metal-Amide Polycrystalline Materials Feature Blueshifted Energy yet Unchanged Kinetics of Emission, Chem. Mater., 2020, 32, 4743–4753.
19. J. Hossain, R. Akhtar and S. Khan, Luminescent coinage metal complexes of carbenes, Polyhedron, 2021, 201, 115151.
20. Q. Gu, F. Chotard, J. Eng, A.-P. M. Reponen, I. J. Vitorica-Yrezabal, A. W. Woodward, T. J. Penfold, D. Credgington, M. Bochmann and A. S. Romanov, Excited state lifetime modulation by twisted and tilted molecular design in carbene-metal-amide photoemitters, Chem. Mater., 2022, 34(16), 7526–7542.
21. A. Ying and S. Gong, A Rising Star: Luminescent Carbene-Metal-Amide Complexes, Chem. – Eur. J., 2023, 29, e202301885.
22. H. Uoyama, K. Goushi, K. Shizu, H. Nomura and C. Adachi, Highly efficient organic light-emitting diodes from delayed fluorescence, Nature, 2012, 492, 234–238.
23. Y. Liu, C. Li, Z. Ren, S. Yan and M. Bryce, All-organic thermally activated delayed fluorescence materials for organic light-emitting diodes, Nat. Rev. Mater., 2018, 3, 18020.
24. N. L. Phuoc, A. C. Brannan, A. S. Romanov and M. Linnolahti, Tailoring Carbene–Metal–Amides for Thermally Activated Delayed Fluorescence: A Computationally Guided Study on the Effect of Cyclic (Alkyl)(amino)carbenes, Molecules, 2023, 28, 4398.
25. J. Callaway and N. H. March, Density Functional Methods: Theory and Applications, Solid State Phys., 1984, 38, 135–221.
26. R. G. Parr and W. Yang, Density-Functional Theory of Atoms and Molecules, Oxford University Press, Oxford, 1989.
27. F. Furche and D. Rappoport, Density Functional Methods for Excited States: Equilibrium Structure and Electronic Spectra. in Computational Photochemistry, ed. M. Olivuccim, Elsevier, Amsterdam, Netherlands, 2005, 16, 93–128.
28. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb and J. R. Cheeseman and et al., Gaussian 16, Revision A. 03, Gaussian Inc., Wallingford CT, USA, 2016.
29. H. S. Yu, X. He, S. L. Li and D. G. Truhlar, MN15: A Kohn–Sham global-hybrid exchange–correlation density functional with broad accuracy for multi-reference and single-reference systems and noncovalent interactions, Chem. Sci., 2016, 7, 5032–5051.
30. F. Weigend, M. Häser, H. Patzelt and R. Ahlrichs, RI-MP2: optimized auxiliary basis sets and demonstration of efficiency, Chem. Phys. Lett., 1998, 294, 143–152.
31. F. Weigend and R. Ahlrichs, Balanced basis sets of split valence, triple zeta valence and quadruple zeta valence quality for H to Rn: Design and assessment of accuracy, Phys. Chem. Chem. Phys., 2005, 7, 3297–3305.
32. F. Neese, Software update: the ORCA program system, version 5.0, WIRES Comput. Molec. Sci., 2022, 12(1), e1606.
33. D. Andrae, U. Häußermann, M. Dolg, H. Stoll and H. Preuß, Energy-Adjusted ab initio Pseudopotentials for the Second and Third Row Transition Elements, Theoret. Chim. Acta, 1990, 77, 123–141.
34. S. Hirata and M. Head-Gordon, Time-dependent density functional theory within the Tamm Dancoff approximation, Chem. Phys. Lett., 1999, 314, 291–299.
35. A. Dreuw and M. Head-Gordon, Single-Reference ab Initio Methods for the Calculation of Excited States of Large Molecules, Chem. Rev., 2005, 105, 4009–4037.
36. B. Moore, H. Sun, N. Govind, K. Kowalski and J. Autschbach, Charge-Transfer Versus Charge-Transfer-Like Excitations Revisited, J. Chem. Theory Comput., 2015, 11, 3305–3320.
37. J. Feng, L. Yang, A. S. Romanov, J. Ratanapreechachai, A.-P. M. Reponen, S. T. E. Jones, M. Linnolahti, T. J. H. Hele, A. Köhler, H. Bässler, M. Bochmann and D. Credgington, Environmental Control of Triplet Emission in Donor–Bridge–Acceptor Organometallics, Adv. Funct. Mater., 2020, 30, 1908715.
38. T. Lu and F. Chen, Multiwfn: A multifunctional wavefunction analyzer, J. Comput. Chem., 2012, 33, 580–592.
39. T. Chen, L. Zheng, J. Yuan, Z. An, R. Chen, Y. Tao, H. Li, X. Xie and W. Huang, Understanding the Control of Singlet-Triplet Splitting for Organic Exciton Manipulating: A Combined Theoretical and Experimental Approach, Sci. Rep., 2015, 5, 10923.
40. T. Pope, Y. Giret, M. Fsadni, P. Docampo, C. Groves and T. J. Penfold, Modelling the effect of dipole ordering on charge-carrier mobility in organic semiconductors, Org. Electron., 2023, 115, 106760.
41. A.-P. M. Reponen, G. Londi, C. S. B. Matthews, Y. Olivier, A. S. Romanov, N. C. Greenham and A. J. Gillett, Understanding Spin-Triplet Excited States in Carbene-Metal-Amides, Angew. Chem., Int. Ed., 2024, 63, e202402052.
42. Y. Liu, C. Li, Z. Ren, S. Yan and M. Bryce, All-organic thermally activated delayed fluorescence materials for organic light-emitting diodes, Nat. Rev. Mater., 2018, 3, 18020.
43. J. R. Gispert, Coordination Chemistry, Wiley–VCH, Weinheim, Germany, 2008, 483.
44. G. Bao, R. Y. Abe and Y. Akutsu, Bond dissociation energy and thermal stability of energetic materials, J. Therm. Anal. Calorim., 2021, 143, 3439–3445.
45. National Center for Biotechnology Information. PubChem Compound Summary for CID 6736, Skatole. https://pubchem.ncbi.nlm.nih.gov/compound/Skatole. Accessed Apr. 12, 2025.
46. National Center for Biotechnology Information. PubChem Compound Summary for CID 10328, Indoline. https://pubchem.ncbi.nlm.nih.gov/compound/Indoline. Accessed Apr. 12, 2025.
47. National Center for Biotechnology Information. PubChem Compound Summary for CID 5281404, Harman. https://pubchem.ncbi.nlm.nih.gov/compound/Harman. Accessed Apr. 12, 2025.
48. National Center for Biotechnology Information. PubChem Compound Summary for CID 1868, 6-Methoxytryptoline. https://pubchem.ncbi.nlm.nih.gov/compound/6-Methoxytryptoline. Accessed Apr. 12, 2025.
49. A. Endo, M. Ogasawara, A. Takahashi, D. Yokoyama, Y. Kato and C. Adachi, Thermally activated delayed fluorescence from Sn⁴⁺-porphyrin complexes and their application to organic light emitting diodes - A novel mechanism for electroluminescence, Adv. Mater., 2009, 21, 4802–4806.
50. (a) CCDC 2454948: Experimental Crystal Structure Determination.  DOI:10.5517/ccdc.csd.cc2ndkwj; (b) CCDC 2454949: Experimental Crystal Structure Determination.  DOI:10.5517/ccdc.csd.cc2ndkxk.

---