Host‐Guest Doping Enables Room Temperature Phosphorescence from Triarylboranes

Abstract

Revisiting our previous report of RTP in triarylboranes, we found that pure crystalline triarylboranes do not exhibit RTP as initially reported, although all but one purified compound shows RTP in annealed PMMA matrices. Inspired by the impurity responsible for the previously observed RTP, we designed a family of binary host–guest RTP systems constructed by pairing distinct triarylboranes. By tuning the host and guest structures and doping ratios, afterglow visible to the naked eye for up to 8 s can be achieved in crystalline materials under ambient conditions, at doping levels as low as one part per billion. The ultra‐long afterglow makes these materials promising for information encryption, while their high sensitivity to oxygen in non‐annealed PMMA films highlights their potential for sensing applications. Trends in the phosphorescence intensities and lifetimes, dependent on guest molecular structure, are rationalized by computed triplet‐triplet energy transfer rates between the hosts and guests, providing a concrete design strategy for optimizing host‐guest RTP. Overall, this work underscores the critical importance of material purity and advances understanding of the host‐guest interactions governing efficient organic RTP.

While pure triarylboranes show RTP in annealed PMMA films, they do not in crystalline form, contrary to our previous report. However, crystalline host‐guest triarylborane systems show persistent afterglow up to 8 s even at 1 ppb guest loading. Computations identify triplet‐triplet energy transfer as the mechanism governing this efficient RTP.

Introduction

There has been significant interest over the past two decades in purely organic materials that exhibit persistent room temperature phosphorescence (RTP).^{[ 1} , ² , ³ , ⁴ , ⁵ , ⁶ , ^{7 ]} Owing to their low cost, ease of processing, and biological compatibility, these materials have found extensive applications in sensors,^{[ 8} , ⁹ , ¹⁰ , ¹¹ , ^{12 ]} biological imaging,^{[ 13} , ¹⁴ , ¹⁵ , ¹⁶ , ^{17 ]} organic light‐emitting diodes (OLEDs),^{[ 18} , ¹⁹ , ²⁰ , ²¹ , ^{22 ]} X‐ray scintillators,^{[ 23} , ²⁴ , ²⁵ , ²⁶ , ^{27 ]} and security printing technologies.^{[ 28} , ²⁹ , ³⁰ , ³¹ , ^{32 ]} However, in contrast to their metal‐containing counterparts, purely organic materials often exhibit inefficient intersystem crossing (ISC), a process essential for enabling phosphorescence.^{[ 33} , ³⁴ , ³⁵ , ³⁶ , ³⁷ , ^{38 ]} This challenge can be overcome by introducing heavy atoms to enhance spin‐orbit coupling (SOC),^{[ 39} , ⁴⁰ , ⁴¹ , ⁴² , ^{43 ]} or by incorporating functional groups with lone pairs, such as carbonyls, to accelerate ISC according to El‐Sayed's rule.^{[ 44} , ⁴⁵ , ⁴⁶ , ⁴⁷ , ⁴⁸ , ⁴⁹ , ^{50 ]} Our previous theoretical work demonstrated that the introduction of empty p‐orbitals on boron can similarly accelerate ISC in triarylboranes via (σ, B p)→(π, B p) transitions.^{[ 51 ]} Furthermore, strategies such as crystal engineering, host–guest doping, polymerization, and chemical crosslinking have been employed to suppress molecular vibrations and non‐radiative decay, thereby improving phosphorescence efficiency.^{[ 52} , ⁵³ , ⁵⁴ , ⁵⁵ , ⁵⁶ , ^{57 ]} As a result, numerous organic RTP materials have been successfully developed using these approaches.^{[ 58} , ⁵⁹ , ⁶⁰ , ⁶¹ , ⁶² , ^{63 ]}

In the rapidly advancing field of organic RTP materials, three‐coordinate boron motifs have attracted particular attention^{[ 7} , ^{38 ]} because the empty p‐orbital at the boron center can conjugate with adjacent π‐systems, yielding distinctive electronic and optical properties.^{[ 64} , ⁶⁵ , ⁶⁶ , ⁶⁷ , ⁶⁸ , ⁶⁹ , ⁷⁰ , ⁷¹ , ⁷² , ⁷³ , ⁷⁴ , ⁷⁵ , ⁷⁶ , ⁷⁷ , ⁷⁸ , ⁷⁹ , ⁸⁰ , ⁸¹ , ⁸² , ⁸³ , ^{84 ]} These properties can be precisely tuned by modifying the substituents on the aryl rings connected to the boron atom.^{[ 64} , ⁶⁵ , ⁶⁶ , ⁶⁷ , ⁶⁸ , ⁶⁹ , ⁷⁰ , ⁷¹ , ⁷² , ⁷³ , ⁷⁴ , ⁷⁵ , ⁷⁶ , ⁷⁷ , ⁷⁸ , ⁷⁹ , ⁸⁰ , ⁸¹ , ⁸² , ⁸³ , ^{84 ]} In 2020, we reported that certain triarylboranes exhibited persistent room temperature phosphorescence in their pure crystalline state.^{[ 51 ]} However, recent studies have increasingly shown that trace impurities can induce RTP in materials that do not exhibit RTP in their pure state.^{[ 85} , ⁸⁶ , ⁸⁷ , ⁸⁸ , ⁸⁹ , ⁹⁰ , ^{91 ]} These impurities may originate from commercial starting materials or from byproducts or intermediates formed during synthesis that are not completely removed during purification. For example, in 2020, Chen et al. demonstrated that commercial sources of carbazole contain the carbazole isomer 1H‐benzo[f]indole, and this trace impurity is responsible for the persistent afterglow observed in numerous carbazole‐based phosphors.^{[ 92 ]} In 2021, Zhang and coworkers demonstrated that in phthalimide‐based hosts, a side product formed via reaction with DMF can produce RTP, even at concentrations on the order of parts per billion (ppb).^{[ 93 ]} In the same year, Ma et al. showed that trace impurities arising from the reaction of imidazole with 4‐bromofluorobenzene in DMF at 100 °C profoundly influenced the phosphorescence of the final product, even at concentrations as low as 0.01 mol%.^{[ 94 ]} In 2024, one of our groups reported that a bis(triarylamine) impurity generated during synthesis was responsible for the greenish yellow phosphorescence of a triarylamine boronic ester,^{[ 95 ]} which had previously been reported to exhibit RTP as a pure compound.^{[ 96 ]} We also re‐evaluated another study reporting RTP from other arylboronic esters.^{[ 97 ]} Careful purification revealed that the observed afterglow originated from trace impurities^{[ 98 ]} rather than the pure compounds. In another example, Marder and co‐workers clarified that pure boric acid does not exhibit RTP,^{[ 99 ]} contrary to an earlier report.^{[ 100 ]} Others have proposed that formation of a B‐O‐O‐B impurity^{[ 101 ]} or trapping defects induced by thermal processing^{[ 89 ]} can be responsible for RTP in boric acid, although the original authors attribute it to a clusterization‐triggered emission (CTE) mechanism.^{[ 102 ]}

These examples illustrate the ubiquity of impurity‐induced persistent phosphorescence in solid‐state organic photoluminescent materials and motivated us to re‐examine our own previous report of RTP from pure triarylboranes. Herein, we re‐synthesized two compounds, DXB‐Ph and DXB‐Xyl (Figure 1), that we had previously reported to exhibit RTP as pure crystalline materials.^{[ 51 ]} We found that the phosphorescence intensity of these compounds in the crystalline state gradually decreased with successive purification until vanishing completely, indicating that the previously reported persistent afterglow was impurity‐induced. We identified one impurity, [1,1′‐biphenyl]‐4‐ylbis(2,6‐dimethylphenyl)borane (DXB‐p‐Bp), which we established as the species responsible for the green RTP originally attributed to pure DXB‐Ph (Figures 1, and S1–S4). To investigate RTP from triarylboranes systematically,^{[ 103} , ¹⁰⁴ , ^{105 ]} we utilized the previously reported compounds as hosts, and synthesized additional mesityl‐substituted analogues DMB‐Ph and DMB‐Mes (i.e., trimesitylborane),^{[ 106} , ¹⁰⁷ , ^{108 ]} finding DMB‐Ph to be particularly suitable for structure‐property correlation studies. Additional guest molecules DMB‐m‐Bp and DMB‐o‐Bp (Figure S5), substituted at different positions of the biphenyl moiety, were prepared for comparison with the original para‐substituted biphenyl impurity.^{[ 109 ]} Inspired by recent work from Li^{[ 110 ]} and Thilagar,^{[ 111 ]} we also designed and synthesized a series of naphthalene‐substituted guests (Figure 1). Remarkably, strong greenish‐yellow afterglow lasting up to 8 s was observed upon doping the 1‐naphthalene‐substituted guests (DXB‐1‐Nap, DMB‐1‐Nap) into their corresponding hosts. Through detailed density functional theory (DFT) analysis^{[ 112 ]} and the application of classical Marcus theory to triplet‐triplet energy transfer, we investigated how the guest molecular structure influences the rate of host‐guest energy transfer, ultimately controlling the lifetime and intensity of the observed persistent phosphorescence. This work not only emphasizes the critical importance of rigorous purification and characterization in the preparation of organic RTP materials but also provides a framework for the rational design of tunable host‐guest systems that exhibit efficient and long‐lived afterglow.

Figure 1.

Molecular structures involved in a) previously reported impurity‐induced RTP systems and b) this work based on triarylboranes. Hosts are shown in grey and impurities (in part a) or guests (in part b) are shown in blue.

Results and Discussion

All compounds were synthesized in high yield by treating the corresponding arylbromide with n‐BuLi, followed by reaction with one equivalent of bis(2,6‐dimethylphenyl)fluoroborane or dimesitylfluoroborane. In particular, DXB‐Ph was prepared by lithiation of bromobenzene with n‐BuLi, followed by borylation. However, we found that the commercial source of bromobenzene used in this study contained trace amounts of 1,4‐dibromobenzene (Figure S1). As a result, the synthesis of DXB‐Ph also produced the impurity DXB‐p‐Bp via the reaction of bis(2,6‐dimethylphenyl)fluoroborane with lithiated 4‐bromobiphenyl, which is generated from bromobenzene and 1,4‐dibromobenzene reacting with n‐BuLi (Figures S2–S4).

The identification, absorption, and emission spectra of all 14 compounds were recorded in dilute hexane solution (Figure 2, Table 1 and S1). The absorption bands, ranging from 298 to 338 nm, can be attributed to a mixture of π→B and π→π^* transitions. The emission maxima of all compounds fall between 354 and 385 nm, except for DMB‐o‐Bp, which emits at 417 nm. This lower‐energy fluorescence is consistent with the emission spectrum predicted by time‐dependent (TD)‐DFT (Figure S64). In the crystalline state, the emission of all compounds is redshifted relative to their solution spectra, with maxima ranging from 367 to 392 nm, except for DMB‐o‐Bp, which shows a slight blueshift to 412 nm (Figures 2, S6, Tables 1, S1). In both solution and the crystalline state, increasing the number of methyl groups leads to a redshift in the emission wavelength, following the sequence DXB‐Ph < DMB‐Ph < DXB‐Xyl < DMB‐Mes among the four hosts. In addition, the guests containing naphthalene or biphenyl substituents exhibit further redshifts compared to the hosts. For example, the emission maxima of DMB‐1‐Nap, DMB‐2‐Nap, and DMB‐p‐Bp in the solid state are 392, 389 and 385 nm, respectively, whereas DMB‐Ph emits at 373 nm.

Figure 2.

a) Normalized absorption and photoluminescence spectra of pure hosts and guests in hexane (1 × 10⁻⁵ M) and in the crystalline state (λ _ex = 290 nm for solution, λ _ex = 310 nm for solid‐state). b) Photographs of all hosts and guests in the crystalline state under 365 nm UV irradiation under ambient conditions.

Table 1.

Photophysical data for pure DMB‐Ph, DMB‐p‐Bp, DMB‐1‐Nap, DMB‐2‐Nap, and their corresponding doped systems (0.0001, 0.001, 0.01, 0.1, 1.0, and 10.0 mol%).

	State	Φ PL %	Φ p %	λ f nm	τ f ns	λ DF nm	τ DF ms	λ p nm	τ p ms
DMB‐Ph (RT)	Hex. a)	3.7	–	360	1.2	–	–	–	–
	Crystal b)	4.6	–	373	1.1	–	–	–	–
DMB‐Ph (77 K)	Tol. c)	–	–	358	–	–	–	414, 437	1031
	Crystal c)	–	–	365	–	–	–	457	2220
DMB‐p‐Bp (RT)	Hex. a)	14.1	–	366	1.4	–	–	–	–
	Crystal b)	20.8	–	385	1.4	–	–	–	–
DMB‐p‐Bp (77 K)	Tol. c)	–	–	361	–	–	–	464, 496	3249
	Crystal c)	–	–	384	–	384	48	488, 522	1242
DMB‐1‐Nap (RT)	Hex. a)	19.6	–	387	1.9	–	–	–	–
	Crystal b)	14.6	–	392	1.5	–	–	–	–
DMB‐1‐Nap (77 K)	Tol. c)	–	–	373	–	–	–	496, 535, 580	1307
	Crystal c)	–	–	394	–	395	6	532, 570	355
DMB‐2‐Nap (RT)	Hex. a)	17.3	–	366, 383	3.3	–	–	–	–
	Crystal b)	6.5	–	389	3.0	–	–	–	–
DMB‐2‐Nap (77 K)	Tol. c)	–	–	372, 389	–	–	–	491, 530, 575	1842
	Crystal c)	–	–	376, 395	–	376, 395	18	555	528
DMB‐p‐Bp/ DMB‐Ph (RT, air)	1:106	4.9	n/d	375	1.1	378	n/d	518, 556	n/d
	1:105	5.3	n/d	373	1.1	374	n/d	518, 558	n/d
	1:104	4.8	0.3	374	1.1	373	1.8	481; 514	43
	1:103	5.3	0.4	374	1.1	373	5.0	481, 514	154
	1:102	6.4	0.6	376	1.2	375	26.1	481, 514	185
	1:10	7.7	0.7	380	1.3	379	17.3	481, 514	177
DMB‐1‐Nap/ DMB‐Ph (RT, air)	1:106	4.6	n/d	368	1.2	n/d	n/d	515, 555, 600	748
	1:105	5.3	n/d	375	1.4	389	n/d	515, 555, 600	748
	1:104	6.5	0.8	380	1.4	390	9.0	515, 555, 600	751
	1:103	18.0	2.2	386	1.7	391	8.7	515, 555, 600	751
	1:102	25.3	4.4	391	1.7	393	8.8	515, 555, 600	756
	1:10	31.0	4.6	396	1.7	394	3.8	515, 555, 600	730
DMB‐2‐Nap/ DMB‐Ph (RT, air)	1:106	4.3	n/d	374	1.3	n/d	n/d	542	n/d
	1:105	4.3	n/d	376	1.3	n/d	n/d	538	570
	1:104	4.1	0.3	378	1.3	n/d	n/d	506, 540	527
	1:103	4.4	0.4	382	1.6	n/d	n/d	506, 542, 582	516
	1:102	5.9	0.6	386	1.6	n/d	n/d	506, 542, 582	252
	1:10	10.8	0.9	387	2.7	n/d	n/d	505, 542, 582	278

^a) In hexane and

^b) in the crystalline state, no DF or phosphorescence were detected at room temperature (RT).

^c) In toluene at 77 K, fluorescence, phosphorescence and its lifetime were measured; n/d = not detected.

Importantly, no delayed emission was observed for any of the pure hosts or guests in the crystalline state under ambient conditions (Figures S7–S9). Because the previously observed RTP properties of DXB‐Xyl and DMB‐Ph were in question,^{[ 51 ]} any compounds in the present study exhibiting RTP were subjected to repeated purification by column chromatography until phosphorescence was no longer detectable in the crystalline state (See “General purification procedure” in the Supporting Information). This procedure confirmed that the previously reported RTP in the crystalline state originated from trace impurities at concentrations below the detection limit of NMR spectroscopy.

Photophysical measurements were also performed for all compounds in toluene at 77 K (Figures 3, and S10–S12). At 77 K, the suppression of non‐radiative decay pathways enables both hosts and guests to exhibit strong phosphorescence in toluene, with exceptionally long lifetimes ranging from 0.9 to 3.4 s. Additionally, poly(methyl methacrylate) (PMMA) films doped with pure compounds at 2 wt% were subjected to thermal annealing, and the photophysical properties of these films at room temperature were investigated. Prior to thermal annealing, no phosphorescence was observed for any system in PMMA in air. Oxygen‐quenching and non‐radiative decay are supressed after thermal annealing, leading to phosphorescence lifetimes between 2.9 ms and 1.2 s in air for all pure hosts and guests (Figures S32–S47), except for DMB‐o‐Bp, which exhibits no detectable phosphorescence (Figure S45). This result indicates faster rates of vibrational relaxation in DMB‐o‐Bp relative to the other molecules investigated, consistent with the absence of vibronic structure in its time‐gated photoluminescence spectrum at 77 K in toluene (Figure S12).

Figure 3.

a) Normalized photoluminescence and time‐gated (0.2 ms) spectra (λ _ex = 290 nm) at 77 K in toluene. b) Photographs showing on/off photoluminescence under UV light at 365 nm.

After the pure systems were fully characterized and it was clear that trace impurities could induce RTP in crystalline triarylboranes, binary host‐guest systems were investigated. It was found that dopant ratios significantly influence the bulk photophysical properties of these materials. Three representative host‐guest combinations, DMB‐p‐Bp/DMB‐Ph, DMB‐1‐Nap/DMB‐Ph, and DMB‐2‐Nap/DMB‐Ph, were selected for systematic study at six different ratios ranging from 0.0001 to 10.0 mol% (Figure 4, Table 1). Throughout this work, the notation “guest/host” is used, with the percentage denoting the molar fraction of guest molecules. Powder X‐ray diffraction (PXRD) experiments were conducted on pure DMB‐Ph, DMB‐1‐Nap, and three binary DMB‐1‐Nap/DMB‐Ph systems at 0.1, 1.0, and 10.0 mol% doping levels. This yielded nearly identical diffraction patterns (Figure S48) for pure DMB‐Ph and the three binary systems, indicating that the doping ratio has a negligible impact on the crystal packing of the host. In the photoluminescent (PL) spectra of the doped systems, the fluorescence emission exhibits a redshift as the guest concentration increases, consistent with enhanced fluorescence from the guest molecules.

Figure 4.

a), b), and c) Normalized steady state photoluminescence spectra; d), e), and f) Time gated (0.2 ms delay) emission spectra, g), h), i) Phosphorescence decays; and j), k), and l) Photographs under 365 nm lamp on/off of doped systems DMB‐1‐Nap/DMB‐Ph, DMB‐2‐Nap/DMB‐Ph, and DMB‐p‐Bp/DMB‐Ph with varying doping ratios (0.0001, 0.001, 0.01, 0.1, 1.0, and 10.0 mol%) in the crystalline state in air at room temperature (λ _ex = 310 nm).

In the DMB‐1‐Nap/DMB‐Ph systems, the steady‐state photoluminescent spectra display a single fluorescence peak around 390 nm, while phosphorescence with vibronic structure appears as peaks at 515, 555, and 600 nm once the doping ratio exceeds 0.01 mol% (Figure 4a). The time‐gated spectra show phosphorescence with a lifetime of approximately 750 ms, largely independent of doping ratio (Figure 4d and g). Remarkably, the generation of phosphorescence is so efficient that it remains detectable at doping ratios as low as 1 ppb (Figure S14). In the DMB‐2‐Nap/DMB‐Ph systems, only fluorescence around 380 nm is observable in the steady‐state spectra, with no distinct phosphorescence peaks (Figure 4b). The time‐gated spectra reveal phosphorescence at 506, 542, and 582 nm, with lifetimes ranging from 252 to 570 ms depending on the doping ratio (Figure 4e and h). Compared with DMB‐1‐Nap, the DMB‐2‐Nap guest produces significantly weaker phosphorescence, as reflected in its lower phosphorescence quantum yields (Table 1). For the DMB‐p‐Bp/DMB‐Ph systems, fluorescence is observed around 375 nm in the steady‐state spectra (Figure 4c). Both delayed fluorescence (DF) and phosphorescence appear in the time‐gated emission spectra when the doping ratio exceeds 0.001 mol% (Figure 4f). Below this concentration, the intensity of both delayed components becomes considerably weaker. As the doping ratio is reduced from 10.0 to 0.01 mol%, the DF lifetime decreases from 17.3 to 1.8 ms, while the phosphorescence lifetime decreases from 177 to 43 ms (Figures 4i, S15). To explore the origin of the DF, temperature‐dependent photophysical measurements were carried out on 1.0 mol% DMB‐p‐Bp/DMB‐Ph (Figure S26). Upon cooling from 180 to 77 K, the DF intensity decreases, consistent with thermally activated delayed fluorescence (TADF). In contrast, for the 1.0 mol% DMB‐1‐Nap/DMB‐Ph system, the DF increases upon cooling from 300 to 77 K (Figure S27). This behavior does not satisfy the conditions for TADF, suggesting that other mechanisms, such as triplet‐triplet annihilation or phosphorescence resonance energy transfer,^{[ 113} , ^{114 ]} may be responsible for the observed DF.

Under 365 nm UV irradiation, systems doped with all three guests show intense persistent luminescence in the crystalline state at room temperature. For example, in the 10.0 mol% DMB‐1‐Nap/DMB‐Ph system in air, the prompt fluorescence vanishes immediately upon switching off the UV light, and an intense greenish afterglow remains visible to the naked eye for 8 s (Figure 4j). Additional systems, including DMB‐Ph‐based systems with 1.0 mol% of other guests (DMB‐m‐Bp, DMB‐o‐Bp) and binary systems based on other hosts (DXB‐Ph, DXB‐Xyl, and DMB‐Mes), were also investigated (Figures S16–S22, Table S2). Among the various guests, naphthalene‐derivatives substituted at the 1‐position produce the most intense and longest‐lived phosphorescence under ambient conditions. The 1.0 mol% DMB‐1‐Nap/DMB‐Ph system exhibits the longest lifetime (756 ms), followed by DXB‐1‐Nap/DXB‐Xyl (331 ms) and DXB‐1‐Nap/DXB‐Ph (256 ms) (Table S2). In the solid state, the fluorescence maxima of pure DMB‐p‐Bp (385 nm), DMB‐1‐Nap (392 nm), and DMB‐2‐Nap (389 nm) exhibit a redshift from their emission in hexane at 366, 387, and 383 nm, respectively (Table 1). This redshift parallels that observed in the doped systems as the concentration of guest molecules is increased. Moreover, the phosphorescence profiles at 77 K for both the doped systems and the corresponding pure guests in toluene are nearly identical (Figures 3a, S26–S28), confirming that the guest molecules are responsible for both fluorescence and phosphorescence in these host‐guest systems.

To confirm that the excited states responsible for RTP originate from the triplet states of the guests, transient absorption spectroscopy was performed on 10 wt% doped PMMA films of DMB‐p‐Bp/DMB‐Ph, DMB‐1‐Nap/DMB‐Ph, and DMB‐2‐Nap/DMB‐Ph (1.0 mol% guest) using a pulsed 355 nm laser and 1 ms delay (Figure 5a–5f). Distinct positive absorption bands with millisecond lifetimes were observed at 426–580 nm for DMB‐p‐Bp/DMB‐Ph and 520–630 nm for DMB‐1‐Nap/DMB‐Ph, confirming that the excited state absorptions originate from long‐lived triplet states rather than short‐lived singlet or charge‐transfer excitations. In the DMB‐2‐Nap/DMB‐Ph (1.0 mol%) system, the PMMA film degenerated when exposed to the laser, resulting in no detectable signal.

Figure 5.

a), d) Transient absorption lifetime density map, b), e) transient absorption spectra, and c), f) phosphorescence decay curves under N₂ of 1.0 mol% DMB‐p‐Bp/DMB‐Ph and DMB‐1‐Nap/DMB‐Ph doped into PMMA thin films at a dopant concentration of 10 wt%. g) Jablonski diagrams of the DMB‐Ph‐based host‐guest systems at the ground state optimized dimer geometry. Singlet (left) and triplet (right) states localized on the host are indicated in grey, while those on the guest molecule are colored.

We next considered the photophysical behavior of the pure host and guest compounds to begin elucidating the mechanism of RTP in these binary systems. All hosts and guests, except DMB‐o‐Bp, exhibit RTP in annealed PMMA films, confirming that their triplet states are accessible via direct excitation of the S₁ state followed by fast ISC. This is consistent with the computational aspects of our previous work, which revealed ISC rates from the S₁ state on the order of 10⁷ s⁻¹ for molecular DXB‐Ph and DXB‐Xyl,^{[ 51 ]} competitive with the radiative decay rate from S₁. Consequently, the absence of RTP in the crystalline state of any of these pure compounds can be attributed to fast vibrational relaxation and triplet quenching. Introducing guest molecules into the bulk host powder must then reduce these modes of non‐radiative decay for the guest, enabling RTP. Furthermore, the significant variation in RTP lifetimes between the host‐guest systems and the unary systems in annealed PMMA films suggests that the host‐guest interaction is critical in the RTP mechanism of the binary systems (Figure 6d).

Figure 6.

a) Proposed mechanism of RTP in the investigated triarylborane‐based systems. In binary host‐guest systems, photoexcitation (A) of the host is followed by intersystem crossing (ISC) and Dexter energy transfer (EnT) to the guest T₁ state. Reverse EnT (rEnT) to the host acts as a deactivation pathway for the guest T₁ state, influencing the phosphorescence (P) lifetime. In unary systems, RTP proceeds via direct excitation of the phosphorescent triarylborane followed by ISC. Non‐radiative decay (wavy lines), such as vibrational relaxation and oxygen‐ or self‐quenching, are suppressed in crystalline triarylborane matrices and annealed PMMA for all guests except DMB‐o‐Bp. b) Hole‐electron (light‐dark) diagrams for the host and guest T ₁ states of the DMB‐1‐Nap/DMB‐Ph dimer at their respective minimum geometries. c) Potential energy surfaces between the minimized dimer geometries of the host and guest T ₁ states. The activation barrier for energy transfer from the host to guest is indicated in grey, while the reverse barrier is indicated in color. DMB‐o‐Bp/DMB‐Ph is excluded since no RTP lifetime was observed. d) Correlation between the RTP lifetimes of the 1 wt.% host‐guest systems and the pure guests at 2 wt. % in annealed PMMA, indicating distinct RTP mechanisms for the unary and binary systems. e) Relationships between the inverse of the RTP lifetime (1/τ) and the Arrhenius‐type rate constants for Dexter energy transfer (k_EnT) based on the forward (∆G ^‡ _for) and reverse (∆G ^‡ _rev) activation barriers shown in c).

The energy transfer mechanism between the interacting host and guest must then be elucidated. In some RTP systems, energy transfer is proposed to proceed through a charge‐separated intermediate, such as an exciplex or radical pair.^{[ 115} , ¹¹⁶ , ^{117 ]} For the radical pair mechanism to operate, the formation of a fully charge‐separated state between the host and guest is critical.^{[ 118 ]} However, the EPR spectra of 1.0 mol% DMB‐1‐Nap/DMB‐Ph, DMB‐2‐Nap/DMB‐Ph, and DMB‐p‐Bp/DMB‐Ph reveal no detectable signals under irradiation (Figures S49–S53), inconsistent with a radical charge transfer mechanism. Moreover, a charge separation process would require the LUMO energy level of the guest to be higher than that of the host.^{[ 115} , ^{119 ]} Our calculated LUMO energy levels for the guests DMB‐1‐Nap, DMB‐2‐Nap, and DMB‐p‐Bp are −1.12, −1.15, and −1.17 eV, respectively, which are not conducive to a charge separation process in combination with the higher DMB‐Ph LUMO energy level of ‐0.98 eV (Table S51). Alternatively, for an exciplex‐mediated mechanism, EPR activity is not expected, but computational analysis of a dimer of the host and guest should reveal a CT state involving the HOMO on the donor molecule and the LUMO on the acceptor molecule.^{[ 120 ]} DFT analysis of the first 9 triplet states for the lowest energy dimer conformations of DMB‐1‐Nap, DMB‐2‐Nap, and DMB‐p‐Nap with DMB‐Ph reveals no bimolecular charge transfer states for these systems (Figure S66).

This leaves the most plausible mechanism for RTP in these binary systems to be direct Dexter triplet‐triplet energy transfer (Figure 6a).^{[ 121} , ^{122 ]} Quantitative modeling of Dexter energy transfer is notoriously difficult due to the computationally demanding electronic coupling term in the semi‐classical formulation of Marcus theory. This often limits analysis of this mechanism to the qualitative matching of energy levels between the donor and acceptor molecules. In our systems, the ground state Jablonski diagrams (Figure 5g) indicate a very large energy gap between T ₁ of the host and T ₁ of the guest, particularly in the DMB‐1‐Nap/DMB‐Ph and DMB‐2‐Nap/DMB‐Ph systems, which would predict unfavorable energy transfer for our most efficient systems. However, the rate cof triplet‐triplet Dexter energy transfer (k _EnT) has recently been shown to be successfully described by a purely classical variant of Marcus theory,^{[ 123 ]} leading to an Arrhenius‐type equation (Equation 1), where A is a pre‐exponential factor, ∆G ^‡ is the activation barrier, R is the gas constant, and T is the temperature. This classical framework has been successfully applied to describe exciplex‐mediated electron exchange in host‐guest RTP systems and Dexter triplet‐triplet energy transfer in photosensitization processes.^{[ 118 ]}

$$k_{EnT}=A{e}^{-\frac{\mathrm{\Delta }{G}^{‡}}{RT}}$$ (1)

Here, we apply this same formalism to our host‐guest systems by optimizing the dimer geometries of each host‐guest pair, where either the host or guest is in its T ₁ state and the opposing molecule is in its ground state (S₀) (Figures 6b, S65). The OT‐S‐LC‐ωhPBE/6–31G(d)/GD3BJ/IEFPCM(hexane) level of theory was employed as it provides close agreement with the experimental photophysical properties of the individual host and guest molecules (Figure S64). Grimme's D3 dispersion correction was also applied to account for any non‐covalent interactions between the host and guest. Interpolated potential energy surfaces between the [T ₁ host / S₀ guest] and [S₀ host / T ₁ guest] geometries were then constructed to approximate the activation barrier for both the forward and reverse energy transfer processes (Figure 6c). Analysis reveals a weak correlation between the inverse of the RTP lifetime and the forward transfer rate, but an excellent correlation with the reverse energy transfer process (Figure 6e). This suggests that the lifetimes of RTP in these host‐guest systems are predominantly governed by the rate of back energy transfer from the guest to the host, which deactivates the phosphorescence pathway of the guest via host non‐radiative decay. This behavior is consistent with other reported binary RTP systems mediated by bi‐directional triplet‐triplet energy transfer, where a low barrier to reverse energy transfer suppresses guest phosphorescence.^{[ 124 ]} As the population of the guest triplet state is controlled by the forward rate of energy transfer, our computational results also correlate with the experimental phosphorescence quantum yields, which depend on triplet population. Among the systems studied, DMB‐1‐Nap exhibits the most intense RTP (Table 1) and has the smallest activation barrier for forward energy transfer. Accordingly, DMB‐m‐Bp is calculated to have the largest forward barrier and, although this system exhibits a measurable phosphorescence lifetime, the phosphorescence quantum yield is too small to be measured within the limits of our spectrophotometer (Table S2). Systems utilizing DXB‐Ph and DXB‐Xyl hosts exhibit better RTP performance than DMB‐Mes‐based systems, while DMB‐Ph‐based systems exhibit the best overall performance. We propose that this may be because the triplet energy levels of DMB‐Ph and its guests are well matched, facilitating efficient forward triplet‐triplet energy transfer.

To demonstrate the potential applications of our systems, we investigated the oxygen sensing performance of doped non‐annealed PMMA films, selecting the 1.0 mol% DMB‐p‐Bp/DMB‐Ph binary system for its high sensitivity to oxygen. No phosphorescence is detected in air for either 1.0 or 10.0 wt% DMB‐p‐Bp/DMB‐Ph (1.0 mol%) doped PMMA films (Figure S29), but both exhibit RTP under N₂, with the 10.0 wt% films exhibiting more intense phosphorescence than the 1.0 wt% films. As the O₂ concentration increases from 0 to 21% in a nitrogen environment, the emission of the 10.0 wt% films shifts from purple to sky blue (Figure 7c). In the absence of oxygen, the films display a persistent green afterglow lasting almost 4 s. This afterglow gradually diminishes with increasing O₂ concentration until it becomes invisible to the naked eye. The ratio of τ ₀/τ increases linearly with increasing oxygen concentration, following a dynamic collisional mechanism, where τ ₀ represents the initial phosphorescence lifetime in a nitrogen atmosphere without oxygen, and τ denotes the phosphorescence lifetime with varying oxygen concentrations (Figure 7b). The ultralong phosphorescence lifetime of the doped systems also indicates promising potential for applications in information encryption. This is demonstrated by painting the 1.0 mol% DMB‐1‐Nap/DMB‐Ph, DMB‐2‐Nap/DMB‐Ph, and DMB‐p‐Bp/DMB‐Ph systems onto weighing paper (Figure 7d). Upon turning off the 365 nm UV light, the papers exhibit a distinct green emission, visible for up to 8 s in the DMB‐1‐Nap/DMB‐Ph system. This highlights the potential utility of this system in security printing and optical data encryption.

Figure 7.

a) Decay profiles b) plots of τ ₀/τ versus oxygen fraction in N₂, and c) photographs of a 10 wt% DMB‐p‐Bp/DMB‐Ph (1.0 mol%) doped non‐annealed PMMA film at various oxygen fractions. d) Illustration of information encryption of DMB‐1‐Nap/DMB‐Ph, DMB‐2‐Nap/DMB‐Ph, and DMB‐p‐Bp/DMB‐Ph (1.0 mol%) on a weighing paper in air under 365 nm irradiation on/off.

Conclusion

In this work, we present a series of efficient binary RTP systems based on triarylborane derivatives. In contrast to our previous findings, we demonstrate that pure triarylboranes do not display RTP in the crystalline state, although intense afterglow emerges in annealed PMMA films for all but one pure compound when non‐radiative decay is suppressed. A trace impurity generated from impure commercial starting materials was identified as the true origin of the previously reported afterglow. Guided by the structure of this impurity, we designed a series of host‐guest systems to study systematically the mechanism of RTP in triarylboranes. Among these systems, the DMB‐1‐Nap/DMB‐Ph (0.1 mol%) system exhibits the longest afterglow, with a lifetime of 752 ms, demonstrating potential for information encryption applications. The DMB‐p‐Bp/DMB‐Ph system, when embedded in PMMA, also displays exceptional oxygen sensitivity, highlighting its suitability for sensing applications. Combining experimental and theoretical insights, we propose that RTP in these binary systems proceeds primarily via ISC from the S ₁ to T ₁ state of the host, followed by direct triplet‐triplet energy transfer to the T ₁ state of the guest. Efficient forward energy transfer leads to high phosphorescence quantum yields, whereas the RTP lifetime is primarily controlled by the rate of back energy transfer from the guest to the host, which acts as a deactivation pathway for phosphorescence from the guest. Systems with small forward and large reverse activation barriers therefore exhibit the most intense and persistent RTP. Overall, this work emphasizes the crucial role of material purity on RTP behavior and provides valuable guidelines for designing multifunctional organic RTP materials with efficient and long‐lived emission.

Conflict of Interests

The authors declare no conflict of interest.

Supporting information

Supporting Information

Data Availability Statement

The data that supports the findings of this study are available in the Supporting Information of this article.