Bodipy Dimer for Enhancing Triplet-Triplet Annihilation Upconversion Performance

Abstract

Triplet-triplet annihilation upconversion (TTA-UC) has considerable potential for emerging applications in bioimaging, optogenetics, photoredox catalysis, solar energy harvesting, etc. Fluoroboron dipyrrole (Bodipy) dyes are an essential type of annihilator in TTA-UC. However, conventional Bodipy dyes generally have large molar extinction coefficients and small Stokes shifts (<20 nm), subjecting them to severe internal filtration effects at high concentrations, and resulting in low upconversion quantum efficiency of TTA-UC systems using Bodipy dyes as annihilators. In this study, a Bodipy dimer (B-2) with large Stokes shifts was synthesized using the strategy of dimerization of an already reported Bodipy annihilator (B-1). Photophysical characterization and theoretical chemical analysis showed that both B-1 and B-2 can couple with the red light-activated photosensitizer PdTPBP to fulfill TTA-UC; however, the higher fluorescence quantum yield of B-2 resulted in a higher upconversion efficiency (η_UC) for PdTPBP/B-2 (10.7%) than for PdTPBP/B-1 (4.0%). This study proposes a new strategy to expand Bodipy Stokes shifts and improve TTA-UC performance, which can facilitate the application of TTA-UC in photonics and biophotonics.

1. Introduction

The photophysical process of photon upconversion converts low-energy photons (long-wavelength light) into high-energy photons (short-wavelength light) [1,2,3,4,5]. Currently, the technology of upconversion luminescence has been extensively utilized in various fields, such as biological imaging [6], 3D printing [7], photocatalysis [8], and solar energy harvesting [9,10]. In particular, triplet-triplet annihilation upconversion (TTA-UC), as the next-generation upconversion material, possesses several distinctive characteristics, such as low excited light power density (~10 mW/cm², sunlight illumination), finely tunable excitation and emission wavelengths, and high upconversion quantum efficiency (η_UC) [11,12,13]. TTA-UC generally consists of two components: photosensitizer (Sen) and annihilator (An). As shown in Figure 1a, in a typical TTA-UC process, the Sen absorbs low-energy incident photons and transits to its singlet excited state (¹Sen*), and subsequently undergoes an intersystem crossing (ISC) to reach its triplet excited state (³Sen*). Then, the energy is transferred from the ³Sen* to the annihilator (An) through the triplet-triplet energy transfer (TTET) process. Finally, two triplet excited states of annihilators (³An*) collision to operate the triplet-triplet annihilation (TTA) process. One annihilator molecule loses energy and returns to its ground state, while the other annihilator molecule transits to its singlet excited state (¹An*) and radiates high-energy upconverted photons [14,15].

Developing an efficient TTA-UC system with excellent η_UC is essential [16,17], in general, which relies on the manipulation of the physiochemical properties of photosensitizers and annihilators [18]. Various types of photosensitizers with intense absorbance and long-lived triplet states, have been developed in prior investigations to significantly improve η_UC [19,20,21,22]. In addition, adjusting the triplet state (T₁) of the annihilator can considerably increase the η_UC via efficiency improvements in the TTET between the photosensitizer and the annihilator. Pioneering studies have been successful in tuning the T₁ of annihilators for superior η_UC, such as diketopyrrolopyrrole (DPP) and tetracene derivatives [23,24]. However, annihilators are often pure organic dyes with no phosphorescence at even low temperatures, making it difficult to determine their T₁ state, so it is difficult to systematically regulate the T₁ energy level of annihilators via a molecular evolution strategy [25]. Boron-dipyrromethene (Bodipy) dyes have been used as model annihilators in TTA-UC because of their high fluorescence quantum yield (Φ_f), robust photostability, and simplicity of chemical functionalization of the molecular structure [26,27,28,29]. For instance, it has been confirmed that the pair of PtTPBP and Bodipy could perform red-to-green or red-to-yellow multicolor photon upconversion [30]. Additionally, perylene and Bodipy moieties were incorporated into the dyad annihilator to widen the ΔE_2T1-S1 and greatly boost η_UC by driving the TTA process [31]. In addition, the pair of TTA-UC (PtTPBP/Bodipy derivatives) successfully creates a ratio-metric nanothermometer with high thermal sensitivity (7.1% K⁻¹) and resolution (0.1 K) that can precisely monitor temperature changes in vivo. This is facilitated by the steric hindrance of the 1,7-dimethyl substituents of Bodipy, which restricts the free rotation of the phenyl moiety at eight sites in Bodipy [32]. However, the small Stokes shift and large molar extinction coefficient of the Bodipy annihilator cause a serious inner-filter effect, which significantly decrease the Φ_f at high concentrations. Furthermore, modulation of the excited state energy level of Bodipy using the π-extension approach could obtain a negative ΔE_2T1-S1 value, such as in 3,5-distyryl Bodipy [26,27,28,29], resulting in a lack of driving force for the TTA process and subsequent decreased upconversion luminescence. The relationship between the T₁ of the Bodipy annihilator and the size of its π-conjugated molecular surface is so poorly understood that it is difficult to regulate their triplet excited states. Therefore, establishing an effective molecular design strategy to regulate the excited states of Bodipy annihilators, specifically the triplet state, remains challenging but intriguing [33,34].

Herein, we describe a dimerization approach at the 2,6-sites of Bodipy to increase the Stokes shift and enhance the performance of the TTA-UC (Figure 1). Compared to the parent Bodipy (B-1), the S₁ of B-2 decreased from 2.43 eV to 2.21 eV, while its T₁ was not reduced significantly after dimerization (1.53 eV vs. 1.52 eV). Calculations based on time-dependent density functional theory (TD-DFT) revealed that the molecular geometry of B-2 tends toward a flat structure in S₁, resulting in a dramatic decrease in S₁ energy level due to an extended π-conjugated surface. In contrast, the two Bodipy moieties tend to have an orthogonal structure in T₁ of B-2, so that the dimerization effect on the T₁ energy level is negligible. Consequently, the double T₁ energy levels of the B-2 are considerably greater than its S₁ (ΔE_2T1-S1 = 0.83 eV), establishing a thermodynamically supported TTA process. Moreover, combining with the red light-absorbing photosensitizer palladium (II) meso-tetraphenyl-tetrabenzoporphyrin (PdTPBP) (Figure 1b), we found that the η_UC of B-2/PdTPBP (10.7%) was significantly higher than that of B-1/PdTPBP (4.0%).

2. Results and Discussion

2.1. Characterization of Photophysical Properties of Annihilators

B-2 was synthesized by a palladium-catalyzed cross-coupling reaction (Scheme 1) [35]. The reaction yield increased from 14% to 61% when compared to a previously reported Bodipy dimerization catalyzed by FeCl₃ (supporting information). B-2 is soluble in common organic solvents, such as toluene, DCM, and EtOAc. Nuclear magnetic resonance (NMR), as well as mass spectrometry (MS), were used to identify the molecular structure.

As shown in Figure 2a, compared to that of B-1, the UV–vis absorption spectrum of B-2 was red-shifted from 503 to 536 nm, and the molar extinction coefficient was as high as 1.39 × 10⁵ M⁻¹ cm⁻¹ (in toluene). The solvent-dependent absorption spectra revealed no change in the absorption profile of B-2, demonstrating that the ground state of B-2 did not undergo charge transfer with the solvent (Figure S1) [36]. The viscosity-dependent absorption spectra of B-1 and B-2 did not exhibit a notable difference (Figure S2).

The fluorescence emission peak of B-2 is 577 nm, which is longer than that of B-1 at 517 nm (Figure 2a). The intersection of the absorption and fluorescence emission spectra shows that the S₁ of B-1 and B-2 are 2.43 and 2.21 eV, respectively, indicating that the S₁ of Bodipy was effectively reduced after dimerization (

Table 1. The photophysical parameters of B-1 and B-2 in toluene.

Compound	λabs a (nm)	ε b	λemc (nm)	Φf1 d (%)	Φf2 e (%)	τf f(ns)	S1 g (eV)	ΔE2T1-S1 h (eV)
B-1	503	0.81	517	72	42	3.52	2.43	0.63
B-2	536	1.39	577	92	86	3.25	2.21	0.83

^a absorption peak; ^b molar extinction coefficient, 10⁵ M⁻¹ cm⁻¹; ^c fluorescence emission peak; ^d fluorescence quantum yield, 10 μM; ^e fluorescence quantum yield, 1000 μM; ^f fluorescence lifetime; ^g single excited energy level was determined as the crossing point of the absorption and fluorescence emission spectra; ^h thermodynamic driving force for TTA, T₁ state energy level was calculated with TD-DFT.

). The absolute Φ_f of B-1 and B-2 at low and high concentrations were then evaluated. The Φ_f of B-1 was 0.72 at a low concentration (10 μM) but 0.42 at a high concentration (1000 μM). However, B-2 retains its Φ_f even at high concentrations (Φ_f = 0.86, 1000 μM) due to its substantial Stokes shift (B-1 vs. B-2, 539 cm⁻¹ vs. 1326 cm⁻¹), which suppresses the inner-filter effect at high concentrations. As demonstrated by the polarity-dependent fluorescence emission spectroscopy, the position and width at half maxima of the emission peak of B-1 changed insignificantly, and the fluorescence intensity was not markedly quenched, indicating that intramolecular charge transfer (ICT) and photoinduced electron transfer (PET) had minor effects on the fluorescent properties of B-1 (Figure S3a) [37]. However, the fluorescence emission of B-2 was significantly repressed in acetonitrile, which is presumably due to the low solubility of B-2 in acetonitrile (Figure S3b) [38]. The viscosity-dependent fluorescence spectroscopy of B-2 demonstrated that the fluorescence emission is not viscosity-dependent (Figure S4). To confirm this, we determined the fluorescence lifetime of B-2 at various viscosities (Figure S5). We did not find that the fluorescence lifetime of B-2 got longer as the viscosity went up. This means that the rotation between the two Bodipy moieties or the 8-site phenyl substitutes does not cause excited-state cone crossings, which would quench fluorescence emission [39].

In addition, we further investigated the redox properties of B-1 and B-2 by cyclic voltammetry versus the Ag/Ag⁺ electrode. The oxidation/reduction potentials of B-1 and B-2 were +0.90/−1.59 V and +0.84/−1.51 V, respectively (Figure S6), demonstrating that Bodipy dimerization does not change the redox potential of the annihilator and, thus, does not result in intramolecular charge separation, which causes fluorescence quenching [40].

Next, the triplet-excited state of B-2 was investigated. Due to the high Φ_f and extremely low triplet state quantum yield of B-2, we were unable to directly observe its phosphorescence emission to determine the T₁ energy level. The T₁ of B-2 was approximated using the triplet sensitization bracketing technique. PdTPBP (T₁ = 1.55 eV) and PtTNP (T₁ = 1.36 eV) are two metalloporphyrins with high phosphorescence quantum yield that have been chosen as triplet energy donors. In the presence of 1000 µM B-1 or B-2, we measured the steady and transient photoluminescence spectra of the triplet energy donor. In the presence of B-2, the phosphorescence of PdTPBP at 800 nm decreased by 50.9% (Figure S7), whereas the phosphorescence intensity of PtTNP remained unchanged at 913 nm (Figure S8a). In addition, the phosphorescence intensity of PdTPBP decreased by 33.1% in the presence of B-1, while that of PtTNP was unaffected. The aforementioned experimental results indicate that the T₁ state energy levels of B-1 and B-2 are between 1.36 eV and 1.55 eV.

To gain a deeper understanding of the excited state of B-2, we calculated its properties using time-dependent density functional theory (TD-DFT). Optimization of the ground state (S₀) configuration of B-2 revealed that the dihedral angle between the two Bodipy moieties is 71° (Figure 3a), suggesting that the steric hindrance of the four methyl substituents in B-2 prevents free rotation between the two Bodipy moieties. The scanning of potential energy surface confirms that the energy of B-2 rises rapidly with the increased co-planarization degree between the two Bodipy moieties (Figure S9). The vertical excitation energy of B-2 corresponds well to its UV–vis absorption spectrum (Table S1), showing that the theoretical model used in the DFT-calculations is valid [41]. In contrast to the optimal conformation of the S₀, the S₁ conformation of B-2 discloses that the dihedral angle between the two Bodipy moieties tends to decrease from 71° to 54° (Figure 3a), indicating that the red-shift of the fluorescence emission of B-2 is associated to the extensive π-conjugated surface between the two Bodipy moieties. Furthermore, the significant differences between S₁ and S₀ in the geometrical configuration result in a large Stokes shift of B-2 [42]. Both Bodipy moieties of B-2 have a frontier orbital electron density population, as calculated by the HOMO and LUMO orbitals of B-2 (Figure 3b). This conclusion is attributable to the enhancement of the π-conjugated surface between the Bodipy moieties resulting from the lowering of the dihedral angle, which is directly associated with the prolonged fluorescence emission wavelength of B-2 [43]. The triplet state of B-2 is then investigated using TD-DFT. As shown in Figure 3a, the dihedral angle between the two Bodipy moieties is 69° in the optimized T₁ conformation of B-2, which is no change in the molecular geometry configuration as compared to S₀. Calculation of the triplet state energy levels of B-2 shows that its T₁ and T₂ states are 1.52 eV and 1.53 eV, respectively. The T₁ and T₂ energy levels are close together due to a lack of π-conjugated electron delocalization between the two Bodipy moieties of B-2, resulting in a degenerate T₁ state [42]. The triplet spin density of B-2 is calculated to be populated in both Bodipy moieties, which further confirmed the abovementioned results (Figure 3c) [41]. In this way, B-1 and B-2 exhibit similar T₁ state energy levels (Table S1). The results of the preceding experimental experiments are consistent with the outcomes of the theoretical chemical calculations.

In the case of Bodipy dyes, both S₁ and T₁ decrease as the π-conjugation surface increases, but T₁ decreases more significantly [26,27,28,29]. This led to a negative ΔE_2T1-S1 value for most Bodipy derivatives, limiting their utilization in TTA-UC [15]. Despite the fact that a dyad annihilator composed of perylene and Bodipy was developed to address this issue, intramolecular charge transfer causes fluorescence quenching and, as a result, reduced TTA-UC efficiency in highly polar solvents [43]. By regulating the molecular geometric configuration of S₁ and T₁, our proposed Bodipy dimerization strategy raises the Stokes shift of the annihilator to prevent dose-mediated upconversion quenching [24].

2.2. TTA-UC Properties of Annihilators

Following that, we chose the red light-absorbing PdTPBP (T₁ energy level = 1.55 eV [44]) as the photosensitizer to investigate the TTA-UC in relation to B-2. Since PdTPBP has a long triplet state lifetime, its own TTA is not conducive to TTET with the annihilator; thus, a low dose of PdTPBP (10 μM) is used. Because of the close dependence of the TTET and TTA processes on the annihilator concentration, we first optimized the B-2 concentration in TTA-UC (Figure S10) [17]. When the concentration of B-2 exceeded 1000 μM, the TTA-UC intensity stopped increasing and even decreased in the concentration interval 1000–1500 μM. This might be because of the self-absorption and fluorescence quenching with the high dose of B-2 [24]. As a result, we measured the η_UC of B-1 and B-2 at 1000 μM, which were determined to be 10.7% and 4.0%, respectively (excited by a 635 nm laser, power intensity = 1267.5 mW/cm²).

As demonstrated in Figure 4a, the TTA-UC peak of B-2 was redshifted to 600 nm compared to B-1. In particular, the TTA-UC spectrum of B-2 exhibits significantly lower intensity at 630 nm compared to the fluorescence spectrum of B-2. This is due to the strong absorption of PdTPBP at this position (Figure S11), indicating the “emission–reabsorption” effect and that the TTA-UC efficiency of PdTPBP/B-2 should be greater than 10.7%. We further evaluated the color of TTA-UC using Commission Internationale de l’Eclairage (CIE) coordinates (Figure 4b). The CIE coordinates of B-1 and B-2 are (0.35, 0.64) and (0.64, 0.36), respectively, indicating that the molecular structure of the annihilator can be finely tuned to produce multicolor TTA-UC. Simultaneously, we observed the TTA-UC colors of B-1 and B-2 in green and yellow, respectively, with the naked eye under low-power red illumination (Figure 4c). The threshold power intensity (I_th) is a critical TTA-UC parameter. The integrated TTA-UC intensity (I_UC) has a linear relationship with the incident light power intensity (I_ex) when I_ex is greater than I_th and a nonlinear quadratic relationship when I_ex is less than I_th. The power-dependent TTA-UC spectra of PdTPBP/B-2 are shown in Figure S12, with a significant increase in I_UC with increasing I_ex. We found a quadratic relationship between the I_UC and the I_ex in the low-power intensity region by logarithmic plotting (Figure 4d). This further demonstrates that the B-2 acts as an annihilator to achieve the red-to-yellow TTA-UC. The I_th of B-2 (53.7 mW cm⁻²) is lower than B-1 (76.9 mW cm⁻²), suggesting that expanding the ΔE_2T1-S1 contributes to the development of TTA-UC pairs with lower I_th, which are desirable for solar energy harvesting, photoredox catalysis, and upconversion bioimaging [6]. In addition, the TTA-UC delayed fluorescence lifetime (τ_DF) of B-1 and B-2 were measured. The τ_DF of B-1 and B-2 were 445.5 μs and 101.4 μs, respectively, which were three orders of magnitude longer than their own τ_f, confirming that this long-lived luminescence was derived from the TTA-UC process [45].

We further measured the TTET efficiency (Φ_TTET) of PdTPBP (10 µM) with B-1 (1000 µM) or B-2 (1000 µM) to gain a better understanding of the TTA-UC procedure. The Φ_TTET of PdTPBP/B-2 (66.9%) is higher than that of PdTPBP/B-1 (49.1%) (

Table 2. The TTA-UC parameters of PdTPBP/B-1 and PdTPBP/B-2 in deaerated toluene.

Compound	ηUC a	ΦTTET b (%)	ηTTA c (%)	Φf d (%)	Ith e	ksv f	kq g	τDF h (µs)	T1 i (eV)
B-1	4.0	49.1	88.0	34.3	76.9	0.57	0.23	445.5	1.53
B-2	10.7	66.9	82.0	60.3	53.7	5.11	2.10	101.4	1.52

^a TTA-UC efficiency at 635 nm CW excitation; ^b triplet-triplet energy transfer efficiency; ^c normalized triplet-triplet annihilation efficiency, 39,570 mW/cm²; ^d absolute fluorescence quantum yields of annihilators (1000 µM) in the presence of PdTPBP (10 µM); ^e threshold power intensity, mW/cm²; ^f Stern–Volmer quenching constant, in 10³ M⁻¹; ^g bimolecular quenching constants, in 10⁷ M⁻¹ s⁻¹; ^h upconversion fluorescence lifetime. ⁱ T₁ state energy levels were calculated with TD-DFT.

). In addition, we calculated the triplet state molecular collision cross-section areas of B-1 and B-2 based on DFT theory [46]. The triplet state collision cross-sections of B-1 and B-2 are 1205.0 Bohr² and 2206.1 Bohr², respectively. The calculation results show that although B-2 has T₁ energy levels similar to B-1, the collision surface of B-2 is greatly extended, which facilitates the Dexter-type triplet energy transfer. Therefore, there is a higher TTET efficiency between PdTPBP and B-2. To further validate our hypothesis, we measured the Stern–Volmer quenching constants (k_sv) of B-1 and B-2 for PdTPBP phosphorescence and then calculated their bimolecular quenching constants (k_q). As shown in Table 2, the k_sv of B-2 is 9.0 times higher than those of B-1, and the k_q is as high as 2.1 × 10⁷ M⁻¹ s⁻¹. The large k_sv of B-2 suggests that the dimerization strategy of Bodipy can promote TTET between photosensitizer and annihilator.

Finally, we measured the normalized triplet annihilation efficiencies (η_TTA) of B-1 and B-2 using a well-established protocol [47]. The η_TTA of PdTPBP/B-1 and PdTPBP/B-2 are 88.0% (39,570 mW/cm²) and 82.0% (39,570 mW/cm²), respectively (Figure S13).

3. Materials and Methods

3.1. Preparation Process for B-2 [35]

A 25 mL three-necked flask was filled with 2-bromo-Bodipy (compound 1) (16.2 mg, 0.05 mmol), X-phos (19.0 mg, 40 mol), bis(pinacolato) diboron (B₂pin₂) (5.1 mg, 0.02 mmol), Cs₂CO₃ (65 mg) 1,4-dioxane (5 mL), and H₂O (200 µL). Then, the mixture was degassed with argon for 10 min, followed by the addition of Pd₂(dba)₃·CHCl₃ (5.2 mg, 5.0 µmol) and another argon degassing for 5 min. After 10 h of reaction at 70 °C, the solvent was evaporated, and the residue was purified using column chromatography on silica with V_hexane/V_DCM = 1:1 to yield 9.5 mg (yield: 61%). ¹H NMR (400 MHz, CDCl₃) (ppm): 7.53–7.43 (m, 6H), 7.34–7.22 (m, 4H), 5.99 (s, 2H), 2.56 (s, 6H), 2.35 (s, 6H), 1.37 (s, 6H), 1.12 (s, 6H). ¹³C NMR (100 MHz, CDCl₃) (ppm): 156.06, 154.71, 143.67, 141.74, 141.25, 135.09, 131.80, 131.28, 129.31, 129.18, 129.07, 128.01, 127.89, 124.78, 121.52, 14.70, 14.44, 13.37, 12.90; MS (MALDI) for B-2 (C₃₈H₃₆B₂F₄N₄) m/z = 646.31 (calculated), 646.31 (observed).

3.2. Measurement of the Fluorescence Quantum Yields (Φ_f) of B-1 and B-2

The literature has previously reported the fluorescence quantum yields of B-1 and B-2 using the relative method with fluorescein as the reference [48]. We utilized the absolute method based on the integrating sphere to determine the fluorescence quantum yields of B-1 and B-2 with greater precision. The absolute method was used to measure the Φ_f in an FLS1000 photoluminescence spectrometer with an integrating sphere and a xenon lamp as the light source. Φ_f is the total number of emitted photons divided by the total number of absorbed photons. We determined the Φ_f of the annihilators at concentration of 10 μM, 1000 μM as well as the annihilators (1000 μM) in the presence of PdTPBP (10 μM). Since the fluorescence emission spectra of B-2 overlaps with the ultraviolet–visible (UV–vis) absorption spectrum of PdTPBP, the Φ_f of the mixture solution B-2/PdTPBP is lower than that of B-2 alone.

3.3. Measurement of Upconversion Efficiency (η_UC)

The η_UC was calculated with the reference Ru(bpy)₃Cl₂·6H₂O, which has a photoluminescence quantum yield (Φ_p) of 0.028 in water [49]. The TTA-UC pairs and Ru(bpy)₃Cl₂·6H₂O were excited using a 635 nm diode laser (39.8 mW) and a 450 nm diode laser (39.8 mW), respectively. We have uniformed the sensitivity of the fluorometer at different wavelengths. The η_UC was calculated using the following equation: eq 1, where η_UC, Φ_f, A_std, A_sam, I_std, I_sam, η_std, and η_sam represent the upconversion efficiency, fluorescence quantum yield of reference, reference absorbance (450 nm), PdTPBP absorbance (635 nm), reference integrated photoluminescence intensity, the integral area of the upconversion spectrum, and the refractive index of H₂O (1.333). Note that the theoretical maximum of η_UC is standardized to be 1 (100%).

$${\eta }_{\mathrm{UC}}=2 \times \Phi f_{ }\times \frac{A_{std}}{A_{sam}}\times \frac{I_{sam}}{I_{std}}\times {\left(\frac{{\eta }_{sam}}{{\eta }_{std}}\right)}^2$$

(1)

3.4. Theoretical Chemical Calculation [50]

All theoretical calculations were performed using the Gaussian 09 program. The ground state geometries of the B-1 and B-2 were optimized using density functional theory (DFT) based on B3LYP/6-31G(d) level. Based on the optimized ground state geometry, the energies of the lowest singlet and triplet excited states were calculated using the TD-DFT method.

4. Conclusions

In conclusion, the new annihilator B-2 was synthesized using the Bodipy dimerization strategy at the 2,6 sites. The η_UC was elevated from 4.0% to 10.7% in comparison to the conventional B-1 annihilator. Through theoretical chemical calculations and spectroscopic characterization, it was confirmed that B-2 not only increases the fluorescence quantum yield but also increases the triplet collisional surface to improve the Φ_TTET. These above features are extremely important for enhancing the η_UC. Thus, this research provides not only a new molecular design strategy for finely regulating the excited state properties of Bodipy-based annihilators but also a new approach for the development of efficient TTA-UC, which will surely promote the use of TTA-UC in photonics and biophotonics fields.