Fast, efficient, narrowband room-temperature phosphorescence from metal-free 1,2-diketones: rational design and the mechanism

Abstract

We report metal-free organic 1,2-diketones that exhibit fast and highly efficient room-temperature phosphorescence (RTP) with high colour purity under various conditions, including solutions. RTP quantum yields reached 38.2% in solution under Ar, 54% in a polymer matrix in air, and 50% in crystalline solids in air. Moreover, the narrowband RTP consistently dominated the steady-state emission, regardless of the molecular environment. Detailed mechanistic studies using ultrafast spectroscopy, single-crystal X-ray structure analysis, and theoretical calculations revealed picosecond intersystem crossing (ISC) followed by RTP from a planar conformation. Notably, the phosphorescence rate constant k_p was unambiguously established as ∼5000 s⁻¹, which is comparable to that of platinum porphyrins (representative heavy-metal phosphor). This inherently large k_p enabled the high-efficiency RTP across diverse molecular environments, thus complementing the streamlined persistent RTP approach. The mechanism behind the photofunction has been elucidated as follows: (1) the large k_p is due to efficient intensity borrowing of the T₁ state from the bright S₃ state, (2) the rapid ISC occurs from the S₁ to the T₃ state because these states are nearly isoenergetic and have a considerable spin–orbit coupling, and (3) the narrowband emission results from the minimal geometry change between the T₁ and S₀ states. Such mechanistic understanding based on molecular orbitals, as well as the structure-RTP property relationship study, highlighted design principles embodied by the diketone planar conformer. The fast RTP strategy enables development of organic phosphors with emissions independent of environmental conditions, thereby offering alternatives to precious-metal based phosphors.

---

Introduction

Room-temperature phosphorescence (RTP) from metal-free organic molecules has been an area of intense research.¹ Although classical RTP materials have been used for diverse applications, including organic light-emitting diodes (OLEDs) and bioimaging, they are mainly precious-metal complexes of Ir or Pt.² Therefore, cost-effective, less-toxic, and sustainable metal-free alternatives are needed. However, intrinsic molecular RTP of metal-free phosphors has not rivalled that of metal complexes.

Organic RTP must overcome several challenges stemming from the absence of heavy-metal atoms. The most severe is the inherently small phosphorescence transition probability (i.e., rate constant k_p), causing poor RTP quantum yields (Φ_p). k_p can be 10⁴–10⁵ s⁻¹ for heavy-metal complexes,³ and only ∼10 s⁻¹ or less for conventional organic compounds.^1a Introducing heavy atoms, such as bromine, iodine, selenium, and tellurium, in conjunction with carbonyl functionalities, is a classical approach for enhancing k_p.⁴ Despite its long history, however, an organic k_p over ∼100 s⁻¹ and/or a Φ_p over 1% in solutions are rarely observed (Fig. 1b, blue filled circles; Fig. S1 and Table S1†).⁵ The use of thiocarbonyls could provide a large k_p in some cases (Fig. 1b, blue open circles).⁶ However, they are associated with high (photo)chemical reactivities and instability,⁷ thereby limiting their high Φ_p in perfluoroalkane solvents.^6b The only practical way to improve Φ_p has been to reduce the nonradiative decay rate constant (k_nr). This approach has provided solid-state phosphors with RTP lasting for subseconds after excitation cessation (persistent RTP or afterglow).⁸ However, it inevitably restricts the organic phosphors to rigid media, such as crystals,⁹ or tailored host–guest systems with strong intermolecular interactions.¹⁰

Fig. 1 Comparison of representative metal-free organic phosphors (blue filled circles), thiocarbonyls (blue open circles), and 1 (orange diamonds) in solution. (a) Molecular structures of 1. (b) Room-temperature phosphorescence quantum yields (Φ_p) vs. lifetimes (τ_p); (c) full-width-at-half-maxima (FWHM) vs. Φ_p. Blue and orange broken lines in (b) represent k_p = 100 and 5000 s⁻¹, respectively, assuming unity intersystem crossing quantum yields.

A promising direction for environmentally independent organic RTP could be the significant increase in k_ps. Complementary to the persistent RTP, fast RTP could have significant potential use in solutions or any non-rigid molecular environment, including biological conditions.¹¹ Moreover, because fast RTP can avoid efficiency loss caused by the accumulation of the triplet excitons, it also could be used in optoelectronics such as OLEDs.¹² However, the quest for fast RTP is facing difficulty so far. In 2022, a selenium-containing molecule (phenoxaselenine) was designed as a candidate for fast RTP; however, its RTP was weak even in a polymer matrix, and only the photoluminescence spectrum and intrinsic phosphorescence rate (ca. 4000 s⁻¹) at 78 K were reported.¹³ Quite recently, organic ionic crystals were reported to exhibit fast RTP with impressive k_p values up to ∼10⁵ s⁻¹.¹⁴ The authors ascribed the large k_p to the external heavy atom effect of two appropriately arranged iodide counter anions. Nonetheless, such effects relied on the crystal packing, and no RTP was observed in solutions. Thus, molecular design for a large k_p that enables RTP in solution is needed.

Low colour purity is another organic RTP challenge. Considering the luminescence of the same quantum efficiency, high-colour-purity emission has a sharp and intense spectral peak that is visible, distinguishable, and vivid. These are essential features for OLED displays and bioimaging.¹⁵ However, organic RTP is usually broad with 80–120 nm full-width at half-maxima (FWHMs) (Fig. 1b, Table S1†). Moreover, because of an inefficient intersystem crossing (ISC) from singlet to triplet states, organic phosphors often exhibit both RTP and fluorescence, further impairing colour purity. As a long-known exception, 1,2-diketones, such as biacetyl and benzil, exhibit narrowband RTP that consists of one main peak with a small FWHM value accompanied by weak vibronic bands.¹⁶ However, the RTP is feeble in solution. Thus, simultaneously achieving a high Φ_p and colour purity in organic RTP remains a tremendous challenge.

Using heteroaromatic 1,2-diketones, we previously developed a series of solid-state mechanoresponsive RTP materials,¹⁷ solvent-free liquid RTP materials,¹⁸ and photoresponsive RTP crystals (Fig. S3†).¹⁹ They exhibited RTP in non-rigid molecular environments, such as amorphous solids or solvent-free liquids, wherein conventional metal-free organic compounds rarely show RTP. Thus, our previous results not only demonstrated unique advantages of the 1,2-diketone-based materials in applications, but also implied large k_ps. However, the fundamental molecular RTP properties remain elusive due to the complexity of the condensed materials. While we have briefly reported some solution-phase RTP properties of thienyl diketones 1a and 1b,^17a further investigations were necessary to evaluate k_p quantitatively (Fig. 1a and S2†). Typically, k_p is derived from experimental Φ_p, ISC quantum yield ϕ_ISC, and RTP lifetime τ_p, expressed as k_p = Φ_p/(ϕ_ISC × τ_p). However, ϕ_ISC was not evaluated in our prior research as we did not focus on k_p. In addition, while we determined Φ_p and τ_p of 1b, the previous protocol did not assure consistent degassing for those measurements (i.e., the extent of oxygen quenching of RTP would be different; Fig. S2†). Consequently, these values cannot be employed to establish k_p. Therefore, the mechanism governing the expected fast RTP remained unexplored, not to mention the absence of the molecular design principle for substantial k_p.

Here, we disclose the inherent molecular RTP properties of the thienyl diketone derivatives 1, exhibiting high-efficiency narrowband RTP based on an exceptionally large k_ps of ∼5000 s⁻¹ (Fig. 1). In particular, 1a exhibited 38.2% Φ_p in cyclohexane. To our knowledge, this is the highest efficiency for metal-free organic molecules in common solvents. Relative to benzil, the k_p was increased by a factor of ∼100 from 39 s⁻¹ to 5300 s⁻¹, enabling a 100-time increase in Φ_p from 0.31% to 38% (Fig. 1a).^16c Moreover, high Φ_p with a narrow ∼30 nm FWHM was observed in solution (Fig. 1b). The high efficiency and colour purity were also observed in various conventional polymer matrices and crystalline solids, indicating that the RTP was an inherent molecular feature. The RTP mechanism and origin were revealed through single-crystal X-ray structure analysis, time-resolved photoluminescence (TRPL) spectroscopy, transient absorption (TA) spectroscopy, time-resolved infrared (TRIR) spectroscopy, and quantum-chemical calculations. Furthermore, structure-RTP property relationship study considering molecular- and natural transition orbitals provided design principles for fast organic RTP. Our work demonstrates the potential of metal-free organic molecular materials to exhibit fast RTP comparable to precious-metal complexes. Complementary to persistent RTP, organic fast RTP is promising not only for application in solution but also in functional condensed materials as we have witnessed in our previous studies (Fig. S3†).

Results and discussion

RTP and related rate constants of 1a

We first examined the RTP properties in solution for the diketone 1a, which exhibits strong yellow emission (Fig. 2a and b). The steady-state photoluminescence spectrum of 1a in cyclohexane had a single, sharp emission peak at 560 nm accompanied by weak vibronic bands at 615 and ∼680 nm.²⁰ The total photoluminescence quantum yield Φ_PL was 38.2% under Ar and the lifetime was 72.7 μs, as determined with a protocol that assures consistent degassing for both measurements (Scheme S2†). TRPL spectra of 1a revealed that the emission emerged just after photoexcitation (within a <100 ps instrumental response function), and exhibited a long lifetime (Fig. S11†). Hence, the emission was phosphorescence, as reported previously for thienyl diketone analogs,¹⁷ and ISC occurred over a timescale less than the instrumental response.

Fig. 2 (a) Steady-state photoluminescence spectra of 1a in cyclohexane (4.4 × 10⁻⁶ M, excited at 368 nm). (b) Photograph of solutions under 365 nm excitation. (c) fs transient absorption spectra of 1a in cyclohexane excited at 355 nm. (d and e) Selected results from the global analysis based on a sequential model; (d) evolution-associated spectra and (e) corresponding concentration kinetics. Coherent artefact signals are omitted for clarity. (f) Schematic summary of the excited-state dynamics in 1a.

To quantify the rapid ISC, we conducted femtosecond TA spectrum (fsTAS) measurement with ∼100 fs time resolution (Fig. 2c). The broad transient absorption over 450–650 nm changed to sharp spectra with a peak at 580 nm in a ∼10 ps timescale. We globally analysed the TAS by using a sequential model assuming two excited species, resulting in successful fitting with the time constant of the transition estimated to be 7.9 ps (Fig. 2d and e and S14†). We also conducted TAS measurements over nanosecond timescales (Fig. S13†), and observed that the TAS shape was unchanged up to the microsecond timescale. The long-lived spectral component was the lowest triplet excited state (T₁ state), populated via ISC with a time constant of <10 ps for 1a in solution (Fig. 2f). Because the ISC outcompeted other relaxation processes, we can reasonably assume that the ISC quantum yield (ϕ_ISC) was unity.

The total photoluminescence quantum yield can be expressed as Φ_PL = Φ_f + Φ_p, where Φ_f is the fluorescence quantum yield. Given that ϕ_ISC ∼ 1, phosphorescence dominates the total emission, and Φ_p ∼ Φ_PL = 0.38. To the best of our knowledge,^21–23 this is a record-breaking RTP efficiency for metal-free organic molecules in common solvents. In addition, with ϕ_ISC ∼ 1 confirmed, Φ_p = ϕ_ISCk_pτ_p ∼ k_pτ_p. For Φ_p = 0.38 and τ_p = 72.7 μs, k_p was derived to be 5300 s⁻¹, which greatly exceeded those of previously reported organic phosphors, except for thiocarbonyl compounds (Fig. 1a and Table S1†). Because of this exceptionally high k_p, the RTP of 1a in solution was visible even in air (Φ_p = 2.6%, Fig. 2a and b, S10, Table S2, and ESI Movie 1†). We note that the estimation of k_p in the literature often assumed ϕ_ISC = 1 and Φ_p ∼ Φ_PL without evaluating the time constant of ISC, and employed Φ_p and τ_p that were determined without assuring consistent degassing. This would be practical for qualitative purposes, but may cause significant errors in the k_p value. Experimental evaluation of ϕ_ISC and assuring consistent degassing are essential for the quantitative evaluation of k_p.

The emission colour purity is also outstanding. Fluorescence was not discernible in the steady-state photoluminescence, and the FWHM of the RTP was only 29 nm for 1a (Fig. 2a and S10†). The RTP contained two other vibronic bands at around 615 and 680 nm; nonetheless, the first band at 560 nm covered 71% of the whole spectral area (Table S3†). This is distinct from other organic phosphors, which have broad and structure-less spectra with a 80–120 nm FWHM (Fig. 1c and Table S1†). Although there is a room for further improvement, the narrowband RTP with weak vibronic bands and the fluorescence-free nature represented a high colour purity emission; its coordinates (0.45, 0.54) are located almost at the edge of the chromaticity diagram of Commission internationale de l'éclairage (CIE) 1931 (Fig. 3b).

Fig. 3 (a) Steady-state photoluminescence (PL) spectra of 1a–1c in cyclohexane (10–4.4 × 10⁻⁶ M, excited at 368 nm) and (b) CIE1931 chromaticity diagram for their PL in cyclohexane. (c) Photophysical properties of 1a–1c in cyclohexane under Ar (10–4.4 × 10⁻⁶ M, excited at 368 nm). Φ_p, RTP quantum yields; FWHM, full-width-at-half-maxima; λ_em, emission maxima; τ_p, lifetimes; k_p and k_nr, phosphorescence and nonradiative rate constants. (d and g) Steady-state PL spectra of (d) 1a–1c@PMMA (5 wt%, excited at 355–368 nm) and (g) a 1c crystal in cyclohexane (1.0 × 10⁻⁵ M, excited at 368 nm). (e and h) Photographs of (e) 1b@PMMA and (h) a 1c crystal under 365 nm excitation. (f) Φ_p in air and the FWHM of 1b-doped polymer films (5 wt%, excited at 350–375 nm). (i) ORTEP drawing of the crystal structure of 1c. Thermal ellipsoids are set at the 50% probability level.

Substituent effect

Next, we investigated the effects of trialkylsilyl substituents on the RTP properties in solution. Trialkylsilyl substituents are known to perturb π-electronic systems through σ–π and/or σ*–π* conjugation (hyperconjugation). While these substituents sometimes improve the fluorescence efficiency, their effects on RTP properties have gained less attention.^9d,24

We introduced triisopropylsilyl (1b) and tributylsilyl groups (1c) onto the thienyl diketone core (Fig. 1a) and examined their properties in cyclohexane solution. The silylated derivatives 1b and 1c provided identical emission spectra with the maxima slightly red-shifted from 560 nm to 568 nm compared to those of 1a (Fig. 3a and S10†). The steady-state emissions were fluorescence-free and maintained a small FWHM (32 nm). This narrowband yellow emission had a CIE 1931 coordinate of (0.49, 051) (Fig. 3b).^2c We also conducted fsTAS measurement for 1b (Fig. S15†) to analyse ISC; the main feature of the TAS was largely the same as that for 1a and the ISC time constants were 0.63 and 2.3 ps. The time constant is even faster than that of 1a and we can assume unity ϕ_ISC for the silylated diketones. Φ_p for 1b and 1c was less than that for 1a, but still 24.6 and 17%, respectively (Fig. 3c). The decrease in Φ_p could be attributed to the doubled k_nr values, from 8500 s⁻¹ for 1a to 15000 s⁻¹ and 22000 s⁻¹ for 1b and 1c, respectively. This was most likely because of the nonradiative decay accelerated by molecular motion involving the silyl groups. In contrast, k_p values were only slightly changed, 5000 and 4500 s⁻¹ for 1b and 1c. These results indicated that the large k_p of 1 basically originated from the thienyl diketone core.

RTP of 1-doped polymer films in air

The huge k_p of 1 enabled efficient RTP in air in the amorphous state using various conventional polymers as matrices (Fig. 3d–f and S20–S24 and Table S4†). The best Φ_p of 54% was achieved at room temperature in air for a poly(methyl methacrylate) (PMMA) film doped with 5-wt% 1a (Fig. 3d). The PMMA films doped with silylated diketones 1b and 1c also provided narrowband PL spectra similar to that of 1a and exhibited a high Φ_p of 48 and 40%, respectively (Fig. 3d and e). These were among the highest reported for polymer-based RTP systems doped with metal-free organic dyes.²⁵ The reported systems with insufficient k_p required a tailor-made approach that involves specific interactions between the polymers and dopants. Hydrogen bonding in poly(vinyl alcohol) was reported to be effective and widely used,^10c–e but these interactions or polymers are difficult to use under highly humid conditions or in water. Because PMMA is resistant to water, the RTP properties of a 1c-doped PMMA film were not spoiled by water, exhibiting the same PL spectrum with a Φ_p of 35% even when submerged in water (Fig. S22†).²⁶

Moreover, phosphorescent films with Φ_p values of 38–18% were obtained by doping 1b into other five conventional polymers (Fig. 3f and S23, S24 and Table S4†). The emission maxima and FWHMs were almost unchanged, with colour purities independent of the polymers. Thus, RTP properties of the films were derived from inherent characteristics of 1.

We would like to emphasize that achieving efficient organic RTP in solution and in polymer matrices is challenging because nonradiative decay is not suppressed under these conditions. In such cases, the fast RTP (large k_p) has a distinct advantage over persistent RTP with small k_p. The same can be said for the RTP in non-crystalline aggregate states, such as the amorphous solid state and solvent-free liquid state. It becomes clear that our previous studies of mechanoresponsive RTP materials,¹⁷ solvent-free liquid RTP materials,¹⁸ and photoresponsive RTP crystals¹⁹ (Fig. S3†) demonstrated the usefulness and applications of fast RTP.

Conformation of RTP-emitting species

Considering that 1 has three successive single bonds in the diketone core, identifying the RTP-emitting conformation was important.^16b–d,27 In our previous study, two distinct conformers of thienyl diketones were identified by single-crystal X-ray structure analysis.^17a The 1a crystal exhibited a trans-planar (TP) conformation with respect to the dicarbonyl moiety, while the 1b crystal exhibited a skew conformation (Fig. S33†). Thus, the conformation in crystals varied based on the substituents due to subtle differences in the intermolecular interactions. Unfortunately, the 1a crystal was nonemissive, likely due to intermolecular electronic interactions. 1a exhibited one-dimensional columnar π-stacking with a 3.481 Å interplanar distance, and the absorption of the crystal had a long tail until over 700 nm (Fig. S34†). On the other hand, the 1b crystal emitted green RTP, which was different from the solution RTP. Therefore, further study was required to nail down the conformation of the yellow RTP-emitting species.

In the present work, we found that crystals of 1c exhibited yellow RTP with a 50% Φ_p in air (Fig. 3g and h). The RTP decayed as a single exponential with a τ_p of 79 μs, and k_p was 6000 s⁻¹ (Fig. S25†). Most importantly, the PL spectrum was almost identical to that of 1c in solution, indicating emission from the same conformer (Fig. 3g). Single-crystal X-ray structure analysis revealed the TP conformation of 1c, with the thienyl diketone core spatially separated from neighbouring molecules by tributylsilyl substituents (Fig. 3i and S32†; no π-stacking, and the interplanar distance was 4.45 Å). Thus, the RTP from the 1c crystal, as well as 1 in solution, was unambiguously assigned to the monomer emission from its TP conformer. Notably, 1c represents a rare example of metal-free organic molecules exhibiting remarkable Φ_p in solution, in a polymer matrix, and in a crystal, with almost identical PL spectra.

To experimentally elucidate conformation dynamics in the excited state, TRIR spectroscopy was performed. Although we could not perform the TRIR measurement on 1a because of poor solubility, we managed to conduct the TRIR measurement on 1b in cyclohexane using a 267 nm wavelength pump light (Fig. S16†). The spectra converged to sharp spectra over 0.74 ps and 2.67 ps timescales, consistent with the time constants extracted from fsTAS of 1b (0.63 and 2.3 ps, Fig. S15†). Note that these time constants were comparable to those observed for 1a in the fsTAS measurement. This indicated that the excited-state dynamics were primarily affected by the thienyl diketone core, while the silyl groups had minor effects. The two timescales were attributed to the conformation change and ISC. More importantly, the converged TRIR spectra were consistent with simulated spectra, assuming a TP conformation in the T₁ state (Fig. S17†). Thus, both the conformation changes and the ISC were ultrafast, and the RTP is confirmed to originate from the TP conformer.

Origin of the rapid ISC

Given that the ISC occurs at the TP conformation, we investigated the molecular origin of the rapid ISC by time-dependent density functional theory (TDDFT) calculations (Fig. 4 and S26†). In principle, ISC is fast when the energy gap is small and the spin–orbit coupling (SOC) matrix element is large.²⁸ In the S₁-optimized TP geometry, the S₁ state was almost isoenergetic to the T₃ state (energy gap <0.01 eV for 1a and 0.04 eV for 1b at the TDA/uCAM-B3LYP-D3/6-311G(d) level of theory).²⁹ Moreover, the S₁–T₃ SOC matrix elements 〈S₁|H_SO|T₃〉 for 1a and 1b were 125 and 112 cm⁻¹, respectively. The substantial SOC matrix elements were consistent with El Sayed's rule; the electronic configurations of the S₁ and T₃ states were (n,π*) and (π,π*), respectively (Fig. S26†). These results strongly supported the ultrafast ISC, which occurred dominantly from the S₁ to T₃ states, followed by internal conversion to the T₁ state.

Fig. 4 Energy diagram of 1a in the S₁-optimized trans-planar (TP) geometry. Energies and the spin–orbit coupling matrix element between the S₁ and T₃ states 〈S₁|H_SO|T₃〉 were calculated at the TDA/uCAM-B3LYP-D3/6-311G(d) level of theory.

Origin of the substantial k_p value

Based on concrete basis that the RTP comes from the TP conformer, we then investigated the origin of the outstanding k_p. Theoretically, it was roughly based on intensity borrowing from the nth excited singlet states (S_n), as given by:^28awhere M_p and μ_Sn–S0 are the T₁–S₀ and S_n–S₀ transition dipole moments at the T₁-geometry, respectively; m_pⁿ contributed to M_p by borrowing intensity from the S_n state; 〈S_n|H_SO|T₁〉 is the SOC matrix element between the S_n and T₁ states. Eqn (1)–(3) provide two insights: (i) m_pⁿ, and hence k_p, increase when S_n states of a large (spin-allowed) μ_Sn–S0 couple with the T₁ state; and (ii) the coupling increases when SOC with energetically close S_n states is effective; 〈S_n|H_SO|T₁〉/ΔE_Sn–T1 can be regarded as the mixing coefficient. It should be emphasized that a large SOC matrix element is insufficient because matching of the three factors in eqn (3)via the same S_n states is the requisite for a significant k_p.

TDDFT calculations for the T₁-optimized TP geometry of 1a yielded k_p = 5400 s⁻¹ (for n = 1–6), which excellently reproduced the experimental value of 5300 s⁻¹. Remarkably, the T₁ state strongly coupled with the S₃ state (Fig. 5a); a large SOC matrix element 〈S₃|H_SO|T₁〉 = 167 cm⁻¹ and a reasonable ΔE_S3–T1 = 1.74 eV = 14 000 cm⁻¹ provided a large mixing coefficient 〈S₃|H_SO|T₁〉/ΔE_S3–T1 of 1.19 × 10⁻². This coefficient indicates that the T₁ state can borrow 1.19% of the transition dipole moment between the S₃ and S₀ states (i.e., μ_S3–S0), which is quite large for phosphorescence. Furthermore, the |μ_S3–S0| was as large as 4.68 D, thus realizing a large |m_p³| = 56 × 10⁻³D. The contribution from m_p³ corresponded to 93% of the total k_p value and is obviously the source of the large k_p (Fig. 6b, vide infra). Similar results were obtained for 1b, confirming that the origin of the substantial k_p lies in the thienyl diketone core (Table S5†).

Fig. 5 (a) Energy diagram of 1a in the T₁-optimized trans-planar (TP) geometry depicting the principal intensity-borrowing. Energies, the transition dipole moment μ_S3–S0, and the spin–orbit coupling matrix element 〈S₃|H_SO|T₁〉 were calculated at the TDA/(u)CAM-B3LYP-D3/6-311G(d) level. (b) Principal natural transition orbitals (NTOs) for S₃–S₀ (left) and T₁–S₀ (right) transitions.

Fig. 6 (a) Chemical structures of 1a, 2a, and 3a, and their Φ_p, estimated k_p, steady-state photoluminescence spectra, and the corresponding FWHM in cyclohexane (1.0 × 10⁻⁵ M, 2.6 × 10⁻³ M, and 1.0 × 10⁻⁴ M, respectively) excited at 368 nm under Ar. k_p was derived as k_p = Φ_p/τ_p, assuming ϕ_ISC = 1. (b) Calculated factors of k_p for 1a–3a. (c) Franck–Condon analysis of the emission spectra of 1a (left) and 2a (right). Huang–Rhys (HR) factors were 0.51 and 2.22, respectively. (d) Bond-elongation ratio between the optimized geometries in the S₀ and T₁ states of 1a and 2a. (e) Natural transition orbitals of the T₁ states for 1a and 2a.

The natural transition orbitals (NTOs) of the S₃ and T₁ states provided an intuitive understanding of the large μ_S3–S0 and 〈S₃|H_SO|T₁〉 (Fig. 5b). First, the S₃ state (S₃–S₀ transition) had a pure (π,π*) configuration, where the NTOs delocalized over the entire molecule in a planar geometry, which was favourable for large μ. On the other hand, the T₁ state had a pure (n,π*) configuration, i.e., it was not contaminated with (π,π*) character. The pure ³(n,π*) configuration enabled SOC with the pure ¹(π,π*) state, consistent with El Sayed's rule.^4a The pure (n,π*)/(π,π*) characters are noteworthy, because excited states with mixed (n,π*) and (π,π*) configurations diminish SOC.³⁰ Most importantly, the hole NTOs of the S₃ and T₁ states around the heavy atoms, especially Br, were mutually perpendicular. This is ideal because the spin–orbit Hamiltonian H_SO involves the orbital angular-momentum operator that rotates the orbitals by 90°.^13,28a Thus, for the hole-constituting atomic orbital of Br, the in-plane σ-symmetric n orbital of T₁ states (Fig. 5b bottom-right) became π-symmetric, producing a significant spatial integral with the out-of-plane p orbital of the S₃ state (Fig. 5b bottom-left). The large spatial integral near the heavy nuclei boosts the heavy-atom effect, resulting in a large 〈S₃|H_SO|T₁〉 = 167 cm⁻¹. Thus, 1 embodies an ideal electronic structure for intensity-borrowing that enhances k_p: (1) a planar π-system that produces large μ; (2) carbonyl groups in the π-plane that joined the pure ¹(π,π*) and pure ³(n,π*) states; and (3) heavy atoms with electrons conjugated to both the π and σ-/n-systems to boost 〈S_n|H_SO|T₁〉.

Structure-RTP property relationship study

The significance of the diketone skeleton for the huge k_p was further evident when comparing 1 with the corresponding bromoaldehyde 2 and diketone without Br atoms 3 (Fig. 6a and b for 1a, 2a, and 3a; Fig. S5† for 1b, 2b, and 3b bearing triisopropylsilyl groups; also see Table S2†). All the compounds exhibited RTP in degassed solutions; however, Φ_p of 1 was much higher than those of 2 and 3 regardless of the silyl substituents (38.2%, 0.08%, and 0.20% for 1a, 2a, and 3a; 24.6%, 0.22%, and 1.2% for 1b, 2b, and 3b, respectively).

The superior Φ_p of 1 was attributed to its large k_p, which was more than one order of magnitude larger than the estimated k_ps of 2 and 3 (Fig. 6a and Table S2†). TDDFT calculations reasonably reproduced this trend; the calculated k_p for 1a, 2a, and 3a was 5400, 9, and 71 s⁻¹, respectively (Fig. S27, S30 and S31†). The small k_p of aldehyde 2 was because of the mismatched 〈S_n|H_SO|T₁〉 and μ_Sn–S0 (Fig. 6b; see Table S4† for 1b, 2b, and 3b). Thus, the T₁ state of 2 had a (π,π*) configuration (Fig. 6e), which coupled strongly to (n,π*) S_n states (e.g., n = 5; 〈S₅|H_SO|T₁〉 was as large as 247 cm⁻¹ for 2a). However, the (n,π*) configuration diminishes μ_Sn–S0 (e.g., |μ_S5–S0| was only 0.22 D for 2a). Consequently, intensity-borrowing was insufficient, resulting in a small k_p. Hence, a large 〈S_n|H_SO|T₁〉 (or a large mixing coefficient, 〈S_n|H_SO|T₁〉/ΔE_Sn–T1) itself is insufficient and the combination with a large μ_Sn–S0 in the same n-th singlet state is essential for achieving a huge k_p.³¹ To meet this requirement, the T₁ state should have a (n,π*) configuration. Indeed, the k_p of Br-free diketone 3a was larger than that of Br-containing aldehyde 2a, because the T₁ state of 3a had a (n,π*) configuration. Thus, the (n,π*) T₁ state prefers coupling with (π,π*) S_n states, whose μ_Sn–S0 can be sizable (e.g., n = 3, |μ_S3–S0| = 5.16 D for 3a; Fig. 6b). However, SOC between these states was not as effective as that for 1a because of the absence of the heavy atom, Br (e.g., 〈S₃|H_SO|T₁〉 = 23 cm⁻¹ for 3a). Therefore, intensity borrowing was less effective, yielding a lower k_p than the brominated diketone 1a.

The side-on spatial arrangement of C–Br and C=O was crucial for the heavy-atom effect because of better mutual alignment of the in-plane Br and O p orbitals. This resulted in large hole coefficients in the T₁ state (Fig. 5b).³² However, the arrangement is energetically unfavourable due to electronic repulsion, because Br and O in the TP geometry of 1 were within the sum of their van der Waals radii according to the X-ray structure (Fig. S32†).^17a,33,34 Indeed, DFT calculations for the aldehyde 2a indicated that the side-on conformation was less stable (by 3 kcal mol⁻¹) for both S₀ and T₁ states than the conformation with oxygen pointing away from the Br (Fig. S28†).³² Interestingly, the TP geometry of 1 involved two-fold intramolecular S⋯O noncovalent chalcogen bonding interactions,³⁵ with interaction energies as large as 7.74 kcal mol⁻¹, according to the natural bonding orbital analysis of 1a.^17a The chalcogen bonds stabilized the TP conformer and increased k_p by forcibly fixing Br in the vicinity of carbonyl oxygens.

Origin of the narrowband emission

Finally, we investigated the origin of the narrowband emission of 1via a comparison of the 1a emission spectra with those of its aldehyde counterpart, 2a (Fig. 6a, c–e; see Fig. S18† for the comparison of 1b with 2b). Diketones 1a and 1b exhibited narrowband emissions (FWHM = 29 and 32 nm) while aldehydes 2a and 2b had broadband emissions (FWHM = 105 and 106 nm).

The bandwidths could be strongly correlated with molecular geometry changes in the ground and excited states, which are quantified using the Huang–Rhys (HR) factor obtained from a Franck–Condon analysis of the spectra (Fig. 6c, S18, and Table S3†).^15,36 The HR factor of diketone 1a (1b) was 0.51 (0.50), while that of aldehyde 2a (2b) was 2.22 (2.39). The smaller HR factors in 1 corresponded to smaller geometry changes in the T₁-to-S₀ transition, which were corroborated by DFT calculations. The results indicated that although diketone 1a and aldehyde 2a retained their planarity in the transition, bond length changes were small for 1a and large for 2a.

We visualized these bond length changes using elongation ratios, defined as (R_T1 − R_S1)/R_S1, where R_T1 and R_S1 represent bond lengths in the T₁-and S₀-optimized geometries; a positive value indicates that the bond is longer in the T₁ state (Fig. 6d). The bond lengths within the thiophene ring of diketone 1a exhibited minimal changes. Conversely, the absolute values of the elongation ratios were large in the case of aldehyde 2a (e.g., the bond length between 3C and 4C changed by 8.4%).

These considerable bond length changes in 2a were attributed to the (π,π*) character of the T₁ state, as evident in the NTOs (Fig. 6e). The electronic transition from bonding to antibonding orbitals significantly affects the bond strength. In contrast, the NTOs of 1a were mainly localized on the 1,2-dicarbonyl moiety, and the electronic transition was nonbonding to antibonding, which led to minor changes in the bond lengths (Fig. 6d). Moreover, the transition in the diketones 1 had no net charge-transfer character because of the centrosymmetric geometry. This was in contrast to most aromatic (mono)carbonyl compounds, where ³(n,π*) states have a charge-transfer character from the carbonyl to the aromatic ring. These features of diketones minimized molecular geometry changes during phosphorescence emission. Therefore, the 1,2-diketone-based phosphor would be a promising platform for efficient narrowband RTP.

Conclusions

Efficient narrowband RTP from metal-free organic 3-bromo-2-thienyl diketones was observed in solutions, amorphous polymer matrices, and crystalline solids. A substantial phosphorescence rate constant of ∼5000 s⁻¹ was experimentally confirmed, which was the key to outstanding RTP quantum yields in solution (38% under Ar) and in polymer films (up to 54% in air). Ultrafast spectroscopy and single-crystal X-ray structure analysis revealed that the fast RTP originated from the diketone planar conformation. Both experimental and theoretical analyses indicated that the planar conformer embodies ideal electronic structures for fast RTP:

1. The planar geometry with a delocalized π-system

2. σ-Symmetric n-orbitals perpendicular to the π-system

3. Heavy atoms conjugated with both π- and n-electron systems

4. (n,π*) configuration of the T₁ state

Fulfilling these points would be a promising design principle for molecules with large μ_Sn–S0 and 〈S_n|H_SO|T₁〉 in the same S_n states, and results in a k_p leap. In addition, the centrosymmetric nonbonding-to-antibonding electronic transition of the 1,2-diketone skeleton was the key to the narrowband emission. The fast RTP (large k_p) is potentially advantageous for various applications, such as phosphorescence OLEDs and lasers, triplet-to-singlet conversion via Förster resonance energy transfer, bioimaging, and theranostics. However, such applications have been monopolized by precious-metal phosphors. We believe that the mechanistic elucidation paves the way for developing fast organic RTP, which could expand applications of metal-free organic materials.

Data availability

All experimental/computational procedures and data related to this article are provided in the ESI.†

Author contributions

Y. T. project administration: lead; supervision: lead; conceptualisation: equal; funding acquisition: lead; investigation: lead; visualisation: lead; writing—original draft: lead; writing—review & editing: lead. K. M. project administration: lead; supervision: lead; conceptualisation: equal; funding acquisition: lead; investigation: supporting; visualisation: supporting; writing—original draft: supporting; writing—review & editing: lead. E. O., Y. O., M. T., and S. K. investigation: equal. M. K. investigation: equal; funding acquisition: supporting; writing—review & editing: supporting. T. E. and K. K. investigation: supporting. K. O. and T. O. resources: lead; supervision: supporting.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by JSPS KAKENHI (grant numbers JP23H03955, JP22H02159, JP20H05676, and JP19K15542). M. K. acknowledges a grant-in-aid for a JSPS Research Fellow (22J12961). Y. T. is grateful to the ENEOS Tonengeneral Research/Development Encouragement & Scholarship Foundation, the Izumi Science and Technology Foundation, the Toyota Physical and Chemical Research Institute, and the Yazaki Memorial Foundation for Science and Technology for the financial support. K. M. acknowledges Sumitomo Basic Science Research Projects, the Kyushu University Q-PIT Support Program for Young Researchers and Doctoral Students, and the Young Researchers Support Project, Faculty of Science, Kyushu University grant numbers 22-A5 (R4). The authors thank Dr Ken-ich Yamashita (Osaka University) for the lifetime measurements and Dr Yuki Kurashige (Kyoto University) and Dr Ryohei Kishi (Osaka University) for fruitful discussions on the quantum chemical calculations. Computations were performed at the Research Center for Computational Science, Okazaki, Japan (Project: 22-IMS-C153 and 23-IMS-C187). The experiments were partially performed at the Analytical Instrument Facility, Graduate School of Science, Osaka University, using research equipment shared in the MEXT project (JPMXS0441200023).

Notes and references

1. (a) S. Hirata, Adv. Opt. Mater., 2017, 5, 1700116; (b) Kenry, C. Chen and B. Liu, Nat. Commun., 2019, 10, 2111.
2. (a) P.-T. Chou and Y. Chi, Chem.–Eur. J., 2007, 13, 380–395; (b) Q. Zhao, C. Huang and F. Li, Chem. Soc. Rev., 2011, 40, 2508–2524; (c) C. Fan and C. Yang, Chem. Soc. Rev., 2014, 43, 6439–6469; (d) C.-P. Tan, Y.-M. Zhong, L.-N. Ji and Z.-W. Mao, Chem. Sci., 2021, 12, 2357–2367.
3. (a) A. Bouskila, B. Drahi, E. Amouyal, I. Sasaki and A. Gaudemer, J. Photochem. Photobiol., A, 2004, 163, 381–388; (b) A. K. Bansal, W. Holzer, A. Penzkofer and T. Tsuboi, Chem. Phys., 2006, 330, 118–129; (c) A. Endo, K. Suzuki, T. Yoshihara, S. Tobita, M. Yahiro and C. Adachi, Chem. Phys. Lett., 2008, 460, 155–157; (d) J. R. Sommer, A. H. Shelton, A. Parthasarathy, I. Ghiviriga, J. R. Reynolds and K. S. Schanze, Chem. Mater., 2011, 23, 5296–5304.
4. (a) N. J. Turro, V. Ramamurthy and J. C. Scaiano, Principles of Molecular Photochemistry: An Introduction, University Science Books, Sausalito, CA, 2009; (b) N. S. e. Konstantin and A. B. Elena, Phys.-Usp., 2005, 48, 231; (c) M. Kasha, J. Chem. Phys., 1952, 20, 71–74; (d) D. S. McClure, J. Chem. Phys., 1949, 17, 905–913; (e) D. R. Kearns and W. A. Case, J. Am. Chem. Soc., 1966, 88, 5087–5097.
5. (a) J. J. Donkerbroek, J. J. Elzas, C. Gooijer, R. W. Frei and N. H. Velthorst, Talanta, 1981, 28, 717–723; (b) J. Xu, A. Takai, Y. Kobayashi and M. Takeuchi, Chem. Commun., 2013, 49, 8447–8449; (c) B. Ventura, A. Bertocco, D. Braga, L. Catalano, S. d'Agostino, F. Grepioni and P. Taddei, J. Phys. Chem. C, 2014, 118, 18646–18658; (d) G. D. Gutierrez, G. T. Sazama, T. Wu, M. A. Baldo and T. M. Swager, J. Org. Chem., 2016, 81, 4789–4796; (e) Z. Yu, Y. Wu, L. Xiao, J. Chen, Q. Liao, J. Yao and H. Fu, J. Am. Chem. Soc., 2017, 139, 6376–6381; (f) H. Shu, H. Li, J. Rao, L. Chen, X. Wang, X. Wu, H. Tian, H. Tong and L. Wang, J. Mater. Chem. C, 2020, 8, 14360–14364.
6. (a) A. Maciejewski and R. P. Steer, Chem. Rev., 1993, 93, 67–98; (b) M. Szymanski, R. P. Steer and A. Maciejewski, Chem. Phys. Lett., 1987, 135, 243–248; (c) M. Szymanski, A. Maciejewski and R. P. Steer, J. Phys. Chem., 1988, 92, 2485–2489.
7. J. D. Coyle, Tetrahedron, 1985, 41, 5393–5425.
8. (a) A. Forni, E. Lucenti, C. Botta and E. Cariati, J. Mater. Chem. C, 2018, 6, 4603–4626; (b) W. Zhao, Z. He and B. Z. Tang, Nat. Rev. Mater., 2020, 5, 869–885; (c) A. D. Nidhankar, Goudappagouda, V. C. Wakchaure and S. S. Babu, Chem. Sci., 2021, 12, 4216–4236.
9. (a) D. B. Clapp, J. Am. Chem. Soc., 1939, 61, 523–524; (b) W. Z. Yuan, X. Y. Shen, H. Zhao, J. W. Y. Lam, L. Tang, P. Lu, C. Wang, Y. Liu, Z. Wang, Q. Zheng, J. Z. Sun, Y. Ma and B. Z. Tang, J. Phys. Chem. C, 2010, 114, 6090–6099; (c) O. Bolton, K. Lee, H. J. Kim, K. Y. Lin and J. Kim, Nat. Chem., 2011, 3, 205–210; (d) M. Shimizu, R. Shigitani, M. Nakatani, K. Kuwabara, Y. Miyake, K. Tajima, H. Sakai and T. Hasobe, J. Phys. Chem. C, 2016, 120, 11631–11639; (e) E. Hamzehpoor and D. F. Perepichka, Angew. Chem., Int. Ed., 2020, 59, 9977–9981; (f) A. Fermi, G. Bergamini, R. Peresutti, E. Marchi, M. Roy, P. Ceroni and M. Gingras, Dyes Pigm., 2014, 110, 113–122; (g) Y. Wen, H. Liu, S. Zhang, Y. Gao, Y. Yan and B. Yang, J. Mater. Chem. C, 2019, 7, 12502–12508.
10. (a) S. Hirata, K. Totani, J. Zhang, T. Yamashita, H. Kaji, S. R. Marder, T. Watanabe and C. Adachi, Adv. Funct. Mater., 2013, 23, 3386–3397; (b) M. S. Kwon, Y. Yu, C. Coburn, A. W. Phillips, K. Chung, A. Shanker, J. Jung, G. Kim, K. Pipe, S. R. Forrest, J. H. Youk, J. Gierschner and J. Kim, Nat. Commun., 2015, 6, 8947; (c) M. S. Kwon, D. Lee, S. Seo, J. Jung and J. Kim, Angew. Chem., Int. Ed., 2014, 53, 11177–11181; (d) Y. Su, S. Z. F. Phua, Y. Li, X. Zhou, D. Jana, G. Liu, W. Q. Lim, W. K. Ong, C. Yang and Y. Zhao, Sci. Adv., 2018, 4, eaas9732; (e) H. Wu, W. Chi, Z. Chen, G. Liu, L. Gu, A. K. Bindra, G. Yang, X. Liu and Y. Zhao, Adv. Funct. Mater., 2019, 29, 1807243; (f) Z.-Y. Zhang, W.-W. Xu, W.-S. Xu, J. Niu, X.-H. Sun and Y. Liu, Angew. Chem., Int. Ed., 2020, 59, 18748–18754.
11. W. Shao and J. Kim, Acc. Chem. Res., 2022, 55, 1573–1585.
12. (a) Z. Zhou, X. Xie, Z. Sun, X. Wang, Z. An and W. Huang, J. Mater. Chem. C, 2023, 11, 3143–3161; (b) H. Imahori, Y. Kobori and H. Kaji, Acc. Mater. Res., 2021, 2, 501–514.
13. W. Shao, H. Jiang, R. Ansari, P. M. Zimmerman and J. Kim, Chem. Sci., 2022, 13, 789–797.
14. I. Partanen, O. Al-Saedy, T. Eskelinen, A. J. Karttunen, J. J. Saarinen, O. Mrózek, A. Steffen, A. Belyaev, P.-T. Chou and I. O. Koshevoy, Angew. Chem., Int. Ed., 2023, 62, e202305108.
15. J. M. Ha, S. H. Hur, A. Pathak, J.-E. Jeong and H. Y. Woo, NPG Asia Mater., 2021, 13, 53.
16. (a) H. L. J. Bäckström and K. Sandros, Acta Chem. Scand., 1960, 14, 48–62; (b) D. J. Morantz and A. J. C. Wright, J. Chem. Phys., 1971, 54, 692–697; (c) L. Flamigni, F. Barigelletti, S. Dellonte and G. Orlandi, J. Photochem., 1983, 21, 237–244; (d) A. K. Singh, D. K. Palit and J. P. Mittal, Chem. Phys. Lett., 2002, 360, 443–452; (e) Y. Tsuboi, K. Okada, S. Ishizaka and N. Kitamura, Anal. Sci., 2005, 21, 303–308; (f) Y. Gong, Y. Tan, H. Li, Y. Zhang, W. Yuan, Y. Zhang, J. Sun and B. Z. Tang, Sci. China: Chem., 2013, 56, 1183–1186.
17. (a) Y. Tani, M. Terasaki, M. Komura and T. Ogawa, J. Mater. Chem. C, 2019, 7, 11926–11931; (b) Y. Tani, M. Komura and T. Ogawa, Chem. Commun., 2020, 56, 6810–6813; (c) Y. Takewaki, T. Ogawa and Y. Tani, Front. Chem., 2022, 9, 812593.
18. M. Komura, T. Ogawa and Y. Tani, Chem. Sci., 2021, 12, 14363–14368.
19. M. Komura, H. Sotome, H. Miyasaka, T. Ogawa and Y. Tani, Chem. Sci., 2023, 14, 5302–5308.
20. Vibronic band is a term for the band due to the transition from the vibrational ground state of an electronic excited state to different vibrational levels of the S₀ (electronic ground) state. We have confirmed this assignment by TRPL measurements, where the whole spectrum decayed at the same rate (Fig. S12†)..
21. In 2014, a family of organic boron complexes was reported to show RTP in solution with a quantum yield over 100% due to singlet fission. However, in 2022, another group evidently pointed out that the RTP was an artifact. They mentioned that “it is obvious that the error is caused by the limit of the dynamic range of the instrument,” and that the compound “exhibits solely a highly intense fluorescence in CH₂Cl₂ with no indication of phosphoresce and dual emission in the room temperature solution.” See ref. 22 for details..
22. (a) M. Koch, K. Perumal, O. Blacque, J. A. Garg, R. Saiganesh, S. Kabilan, K. K. Balasubramanian and K. Venkatesan, Angew. Chem., Int. Ed., 2014, 53, 6378–6382; (b) Y. Chen, C.-H. Wang, T.-C. Chou and P.-T. Chou, Angew. Chem., Int. Ed., 2022, 61, e202109224.
23. During the review process of this manuscript, X. Peng et al. published a report on an efficient organic RTP in solution. The total photoluminescence quantum yield (PLQY) in toluene was reported to be as high as 60% under N₂ and 8% under O₂; however, neither the RTP QY nor the intersystem crossing efficiency was investigated, and the RTP rate constant remains unclear. According to the reported fluorescence lifetimes and PL decay curve, total PLQY under N₂ should include a considerable fluorescence contribution. These results highlight the significance of evaluating the intersystem crossing efficiency. See: X. Peng, P. Zou, J. Zeng, X. Wu, D. Xie, Y. Fu, D. Yang, D. Ma, B. Z. Tang and Z. Zhao, Angew. Chem., Int. Ed., 2024, 63, e202405418.
24. (a) T. Nakashima, M. Shimada, Y. Kurihara, M. Tsuchiya, Y. Yamanoi, E. Nishibori, K. Sugimoto and H. Nishihara, J. Organomet. Chem., 2016, 805, 27–33; (b) M. Shimizu, T. Kinoshita, R. Shigitani, Y. Miyake and K. Tajima, Mater. Chem. Front., 2018, 2, 347–354; (c) S. Hirata, Adv. Sci., 2019, 6, 1900410; (d) M. Shimizu, S. Nagano and T. Kinoshita, Chem.–Eur. J., 2020, 26, 5162–5167.
25. N. Gan, H. Shi, Z. An and W. Huang, Adv. Funct. Mater., 2018, 28, 1802657.
26. D. Lee, O. Bolton, B. C. Kim, J. H. Youk, S. Takayama and J. Kim, J. Am. Chem. Soc., 2013, 135, 6325–6329.
27. (a) Z. He, W. Li, G. Chen, Y. Zhang and W.-Z. Yuan, Chin. Chem. Lett., 2019, 30, 933–936; (b) J. Song, L. Ma, S. Sun, H. Tian and X. Ma, Angew. Chem., Int. Ed., 2022, 61, e202206157; (c) M. Yasui, T. Fujihara, H. Ohtsu, Y. Wada, T. Shimada, Y. Zhu, M. Kawano, K. Hanaya, T. Sugai and S. Higashibayashi, Commun. Chem., 2023, 6, 245; (d) Z. Chen, X. Chen, D. Ma, Z. Mao, J. Zhao and Z. Chi, J. Am. Chem. Soc., 2023, 145, 16748–16759; (e) S. H. Lee, M. S. Valverde Paredes, P. M. Forster and D.-C. Lee, RSC Adv., 2024, 14, 6285–6291.
28. (a) G. Baryshnikov, B. Minaev and H. Ågren, Chem. Rev., 2017, 117, 6500–6537; (b) T. J. Penfold, E. Gindensperger, C. Daniel and C. M. Marian, Chem. Rev., 2018, 118, 6975–7025.
29. M. J. G. Peach and D. J. Tozer, J. Phys. Chem. A, 2012, 116, 9783–9789.
30. H. Ma, Q. Peng, Z. An, W. Huang and Z. Shuai, J. Am. Chem. Soc., 2019, 141, 1010–1015.
31. (a) S. Hirata, J. Phys. Chem. Lett., 2018, 9, 4251–4259; (b) I. Bhattacharjee and S. Hirata, Adv. Mater., 2020, 32, 2001348.
32. S. Sarkar, H. P. Hendrickson, D. Lee, F. DeVine, J. Jung, E. Geva, J. Kim and B. D. Dunietz, J. Phys. Chem. C, 2017, 121, 3771–3777.
33. One might expect the formation of halogen bonding (attractive interactions) in side-on arrangements. However, it is unlikely because the halogen bonding is a highly directional interaction between the lone pair of (in this case) carbonyl O and “sigma-hole” of Br that forms along the C–Br σ* bond. Indeed, there were no donor–acceptor interactions found between Br⋯O according to natural bonding orbital analysis of 1a..
34. S. J. Ang, T. S. Chwee and M. W. Wong, J. Phys. Chem. C, 2018, 122, 12441–12447.
35. (a) D. J. Pascoe, K. B. Ling and S. L. Cockroft, J. Am. Chem. Soc., 2017, 139, 15160–15167; (b) H. Huang, L. Yang, A. Facchetti and T. J. Marks, Chem. Rev., 2017, 117, 10291–10318; (c) R. Gleiter, G. Haberhauer, D. B. Werz, F. Rominger and C. Bleiholder, Chem. Rev., 2018, 118, 2010–2041.
36. M. de Jong, L. Seijo, A. Meijerink and F. T. Rabouw, Phys. Chem. Chem. Phys., 2015, 17, 16959–16969.

---

Footnote

† Electronic supplementary information (ESI) available: Experimental procedures, analytical data, physicochemical properties, computational details, and CIF files for the single-crystal X-ray structure analysis. CCDC 2269866. For ESI and crystallographic data in CIF or other electronic format see DOI: https://doi.org/10.1039/d4sc02841d

---