Supramolecular Annihilator with DPA Parallelly Arranged by Multiple Hydrogen-Bonding Interactions for Enhanced Triplet–Triplet Annihilation Upconversion

The triplet annihilator is a critical component for triplet–triplet annihilation upconversion (TTA-UC); both the photophysical properties of the annihilator and the intermolecular orientation have pivotal effects on the overall efficiency of TTA-UC. Herein, we synthesized two supramolecular annihilators A-1 and A-2 by grafting 9,10-diphenylanthracene (DPA) fragments, which have been widely used as triplet annihilators for TTA-UC, on a macrocyclic host—pillar[5]arenes. In A-1, the orientation of the two DPA units was random, while, in A-2, the two DPA units were pushed to a parallel arrangement by intramolecular hydrogen-bonding interactions. The two compounds showed very similar photophysical properties and host–guest binding affinities toward electron-deficient guests, but showed totally different TTA-UC emissions. The UC quantum yield of A-2 could be optimized to 13.7% when an alkyl ammonia chain-attaching sensitizer S-2 was used, while, for A-1, only 5.1% was achieved. Destroying the hydrogen-bonding interactions by adding MeOH to A-2 significantly decreased the UC emissions, demonstrating that the parallel orientations of the two DPA units contributed greatly to the TTA-UC emissions. These results should be beneficial for annihilator designs and provide a new promising strategy for enhancing TTA-UC emissions.


Introduction
Upconversion (UC) is a phenomenon by which low-energy light is converted into high-energy light.There are several mechanisms to achieve UC emissions, such as multistep excitation of lanthanides ions [1], two-photon absorption [2], and second-harmonic generation [3], among which UC based on triplet-triplet annihilation (TTA) [4][5][6][7] is attracting continuous attention due to its unique advantages of tunable excitation wavelengths and high quantum efficiency.Thus, this has wide application potentials, ranging from photocatalysis [8,9], bio-imaging and bio-sensing [10][11][12][13], and drug release [14,15], to 3D printing [16,17].The triplet sensitizer and annihilator are the essential components for TTA-UC, and a diagram illustrating the mechanism of TTA-UC is shown in Figure S1, in which the triplet-triplet energy transfer (TTET) process from the sensitizer to the annihilator and triplet-triplet annihilation (TTA) between the triplet annihilators are two pivotal processes for TTA-UC, both of which follow a Dexter mechanism.To facilitate the TTET process, photophysical properties of sensitizers [18][19][20] are optimized by either enhancing the visible-light absorption ability and/or extending the lifetime of their triplet states, and abundant sensitizers ranging from transition-metal complexes to metallic porphyrins and surface-engineered semiconductor nanocrystals were developed for TTA-UC [21][22][23][24].Moreover, strategies for supramolecular pre-assembling of the sensitizers and annihilators in close proximity were also developed to facilitate the TTET process by allowing the TTET event to proceed in an intramolecular fashion instead of an intermolecular one [25][26][27][28][29].
On the other hand, for the TTA process, in which only annihilators participate, there has been far less investigation.As a matter of fact, there are three factors that jointly determine the TTA efficiency: (1) the competition between annihilation and decay for the fate of annihilators' triplet states, (2) the collision probability of the two annihilators' triplet states, and (3) the spin statistical factor (η), which gives the probability that a bright singlet state is formed from a pair of annihilating triplet states.Chemical modification of the annihilators is necessary to enhance the TTA efficiencies.Different types of dimeric annihilators based on DPA [30][31][32][33] and tetracene [34,35], as well as some DPA-based oligomers [36] and dendrimers [37], were synthesized; the triplet lifetime of the annihilator was found to significantly influence the UC emissions [30].Also, achieving TTA in an intramolecular fashion instead of intermolecularly facilitated the TTA process.Significantly, theoretical simulations by Clark et al. showed that parallel chromophores gave higher η than perpendicular chromophores [38].However, the currently developed dimers/oligomer of the annihilators have never achieved parallel arrangement of the annihilator units.
Supramolecular assembly presents an excellent strategy to precisely arrange and control the orientation of molecules [39][40][41][42][43][44][45].Previously, we have demonstrated that chemically tuning the structures of sensitizers, as well as the host-guest complexation, could significantly facilitate the TTET process [46,47], and, by carefully regulating the aggregation state of the annihilator, the TTA process was selectively manipulated without disturbing the TTET process [48,49].Moreover, by self-assembling the annihilators on the edge of an inorganic clay nanosheet through electrostatic attraction, a UC quantum yield (Φ UC ) record in the solid state was achieved [50].These results demonstrated that, in addition to the photophysical properties of the UC components, the orientation of molecules also had an important effect on the TTA-UC efficiencies.Thus, herein, we designed two novel supramolecular annihilators, A-1 and A-2, to confirm the important effect of the molecular orientation on the TTA process.It was found that, along with the facilitated TTET and TTA process, due to the close proximity of the sensitizer and annihilator, the parallel orientation of the DPA units in A-2 greatly enhanced the TTA process, with UC quantum yield enhanced from 5.1% for A-1 to 13.7% for A-2.

Synthesis
The syntheses of A-1 and A-2 are outlined in Scheme 1, while the synthesis of the sensitizers is shown in Scheme S1.The intermediate M2 was prepared by a palladium-catalyzed Suzuki cross-coupling of 9,10-dibromoanthracene with (4-(hydroxymethyl) phenyl) boronic acid, followed by a bromination of M1 with CBr 4 .The intermediate P [5]OH was prepared by following a reported procedure, and the detailed description of the synthesis is displayed in the supporting information.A-Br was then prepared by a nucleophilic substitution reaction of P[5]OH and M2.Methanolysis of A-Br afforded the target product A-1 in good yield of 85.3% over 4 steps.Reaction of A-Br with 1,3-diaminopropane, followed by ammonolysis of cyclohexyl isocyanate with A-NH 2 , gave the other product A-2 in satisfying yield of 67.7% over 5 steps (Scheme 1).The chemical structures of A-1, and A-2, and all the intermediates were explicitly verified by standard methods of 1 H-NMR, 13 C-NMR, and high-resolution MS (Figures S12-S17).

Photophysical Properties of the Annihilators and Sensitizers
The absorption and emission spectra of the supramolecular annihilators A-1 and A-2 are shown in Figure 1a.A-1 and A-2 showed very similar absorptions to that of DPA.The molar extinction coefficient (ε) of A-1 and A-2 was about twice that of DPA, proportional to the number of DPA units tethering on the pillar [5]arenes, demonstrating that there was no significant interaction between the DPA units in the ground state (Figure 1a).Moreover, A-1 and A-2 showed almost the same emissions, in both wavelength and intensities, and the lifetimes of the fluorescence were also comparable, which were 4.80 ns and 4.73 ns, respectively (Figure S21).The fluorescence quantum yields of A-1 and A-2 decreased a little bit when compared with that of DPA [51], for which the locally concentrated DPA units in A-1 and A-2 was most probably responsible.However, the quantum yields of A-1 and A-2 were comparable with each other, e.g., 65.3% for A-1 and 62.9% for A-2 (Table 1) in toluene, demonstrating that, despite the intramolecular hydrogen-bonding interaction between the carbamido groups in A-2, the excited states' properties of the DPA units were not significantly changed and no π-π stacking between the anthracene core of DPA units was induced, which is highly beneficial for application as the annihilator in TTA-UC.

Photophysical Properties of the Annihilators and Sensitizers
The absorption and emission spectra of the supramolecular annihilators A-1 and A-2 are shown in Figure 1a.A-1 and A-2 showed very similar absorptions to that of DPA.The molar extinction coefficient (ε) of A-1 and A-2 was about twice that of DPA, proportional to the number of DPA units tethering on the pillar [5]arenes, demonstrating that there was no significant interaction between the DPA units in the ground state (Figure 1a).Moreover, A-1 and A-2 showed almost the same emissions, in both wavelength and intensities, and the lifetimes of the fluorescence were also comparable, which were 4.80 ns and 4.73 ns, respectively (Figure S21).The fluorescence quantum yields of A-1 and A-2 decreased a little bit when compared with that of DPA [51], for which the locally concentrated DPA units in A-1 and A-2 was most probably responsible.However, the quantum yields of A-1 and A-2 were comparable with each other, e.g., 65.3% for A-1 and 62.9% for A-2 (Table 1) in toluene, demonstrating that, despite the intramolecular hydrogen-bonding interaction between the carbamido groups in A-2, the excited states' properties of the DPA units were not significantly changed and no π-π stacking between the anthracene core of DPA units was induced, which is highly beneficial for application as the annihilator in TTA-UC.and emission wavelength were bathochromically shifted in S-2 due to the introduction of an amido bond.The maximum phosphorescence wavelength of S-2 red-shifted to 670 nm (1.85 eV, Figure 1b).But, fortunately, the triplet energy level of S-2 was still higher than that of DPA units (1.77 eV) after derivatization, which is essential for the triplet-triplet energy transfer process in TTA-UC.The photophysical parameters of the annihilators and sensitizers are summarized in Table 1.The binding behavior of the supramolecular annihilators A-1 and A-2 toward S-2 was investigated.By gradually adding A-2 to the solutions of S-2, the absorption of S-2 was blue-shifted and continuously increased (Figure 2a); an apparent isoabsorptive point at 590 nm was observed (Figure 2a inset), demonstrating a significant host-guest interaction between A-2 and S-2, and the same was true for A-1 (Figure S23a).More interestingly, the host-guest interaction significantly recovered the phosphorescence of S-2 (Figure S24), for which the inhabitation of the photoinduced electron transfer (PET) from the Pt (II)-salophen complex to the amino cation should be most probably responsible.The binding constants for S-2/A-2 were determined to be 3.6 × 10 4 M −1 (Figure 2b  Pt (II)-salophen complexes were employed as sensitizers for the TTA-UC.The triplet lifetimes of normal Pt (II)-salophen complexes were relatively short, e.g., for S-1, the lifetime was determined to be only 4.4 µs (Figure S22), which was unsatisfactory for an efficient TTET process in TTA-UC in solutions [47].Thus, supramolecular host-guest interactions to position the sensitizer and annihilator in close proximity were highly necessary.Therefore, S-2 was prepared with an alkyl ammonia chain attached on the salophen ligand as the supramolecular binding site for the supramolecular annihilators A-1 and A-2.The lifetime of S-2 was determined to be 2.48 µs, which was comparable with S-1.Despite the nonextended triplet state lifetime, the host-guest complexation between S-2 and A-1/A-2 significantly facilitated the TTET process, which is discussed hereafter.The absorption and emission wavelength were bathochromically shifted in S-2 due to the introduction of an amido bond.The maximum phosphorescence wavelength of S-2 red-shifted to 670 nm (1.85 eV, Figure 1b).But, fortunately, the triplet energy level of S-2 was still higher than that of DPA units (1.77 eV) after derivatization, which is essential for the triplet-triplet energy transfer process in TTA-UC.The photophysical parameters of the annihilators and sensitizers are summarized in Table 1.
The binding behavior of the supramolecular annihilators A-1 and A-2 toward S-2 was investigated.By gradually adding A-2 to the solutions of S-2, the absorption of S-2 was blue-shifted and continuously increased (Figure 2a); an apparent isoabsorptive point at 590 nm was observed (Figure 2a inset), demonstrating a significant host-guest interaction between A-2 and S-2, and the same was true for A-1 (Figure S23a).More interestingly, the host-guest interaction significantly recovered the phosphorescence of S-2 (Figure S24), for which the inhabitation of the photoinduced electron transfer (PET) from the Pt (II)salophen complex to the amino cation should be most probably responsible.The binding constants for S-2/A-2 were determined to be 3.6 × 10 4 M −1 (Figure 2b), while, for S-2/A-1, it was 5.4 × 10 4 M −1 (Figure S23b).We ascribe the slightly smaller binding affinity of A-2 toward S-2 than that of A-1 to the unfavorable conformation of pillar [5]arenes units when intramolecular hydrogen-bonding in A-2 limited the orientation of one benzene ring of the pillar [5]arene.
Molecules 2024, 29, 2203 was 5.4 × 10 4 M −1 (Figure S23b).We ascribe the slightly smaller binding affinity toward S-2 than that of A-1 to the unfavorable conformation of pillar [5]arenes units intramolecular hydrogen-bonding in A-2 limited the orientation of one benzene r the pillar [5]arene.

Applications of A-1 and A-2 as Annihilators for TTA-UC
Upconverted emissions of the annihilators A-1 and A-2 were investigated via ation of the sensitizer S-2 with a 589 nm diode-pump solid-state (DPPS) laser.As in Figure 3a, b, S-2 alone showed red phosphorescence at 600-800 nm, while gra adding A-1 or A-2 from 0.0 μM to 100.0 μM to the S-2 solution, the phosphorescenc 2 gradually decreased, while a new emission at 400-500 nm appeared.The emiss 400-500 nm were essentially linearly dependent on the acceptors' concentrations first stage, while, after adding more than 100.0 μM of A-2, the emission at 400-500 longer increased.The vibration structures of emission spectra were quite similar prompt fluorescences of A-1 and A-2, and irradiating A-1 and A-2 alone with the 5 laser did not result in such emissions.Thus, the emission at 400-500 nm was ascri the sensitized upconverted emission of A-1/A-2 by S-2.Time-resolved emission s showed that the lifetimes of the emissions at 400~500 nm were 247 μs and 200 μs f and A-2, respectively (Figures S25, S26 and 3c), which were overwhelmingly longe the prompt emissions (4.80 ns for A-1; 4.73 ns for A-2), further proving the sen mechanisms.The upconversion intensity of both A-1 and A-2 were significantly inc by increasing the power density (Figures 3d and S27); the double-logarithmic plots integral area of UC emission vs. the power density first followed a quadratic proce then switched to a linear process.The TTA-UC threshold excitation intensity (Ith) of the A-1/S-2 and A-2/S-2 systems were determined to be 1.18 W/cm 2 and 1.08 W respectively, both of which were lower than that of the DPA/S-2 system (Figures 3  and S29), demonstrating that the TTA-UC processes of A-1 and A-2 are more efficien that of DPA.

Applications of A-1 and A-2 as Annihilators for TTA-UC
Upconverted emissions of the annihilators A-1 and A-2 were investigated via irradiation of the sensitizer S-2 with a 589 nm diode-pump solid-state (DPPS) laser.As shown in Figure 3a,b, S-2 alone showed red phosphorescence at 600-800 nm, while gradually adding A-1 or A-2 from 0.0 µM to 100.0 µM to the S-2 solution, the phosphorescence of S-2 gradually decreased, while a new emission at 400-500 nm appeared.The emissions at 400-500 nm were essentially linearly dependent on the acceptors' concentrations at the first stage, while, after adding more than 100.0 µM of A-2, the emission at 400-500 nm no longer increased.The vibration structures of emission spectra were quite similar to the prompt fluorescences of A-1 and A-2, and irradiating A-1 and A-2 alone with the 589 nm laser did not result in such emissions.Thus, the emission at 400-500 nm was ascribed to the sensitized upconverted emission of A-1/A-2 by S-2.Time-resolved emission spectra showed that the lifetimes of the emissions at 400~500 nm were 247 µs and 200 µs for A-1 and A-2, respectively (Figures S25, S26 and 3c), which were overwhelmingly longer than the prompt emissions (4.80 ns for A-1; 4.73 ns for A-2), further proving the sensitized mechanisms.The upconversion intensity of both A-1 and A-2 were significantly increased by increasing the power density (Figures 3d and S27); the double-logarithmic plots of the integral area of UC emission vs. the power density first followed a quadratic process and then switched to a linear process.The TTA-UC threshold excitation intensity (I th ) values of the A-1/S-2 and A-2/S-2 systems were determined to be 1.18 W/cm 2 and 1.08 W/cm 2 , respectively, both of which were lower than that of the DPA/S-2 system (Figures 3d, S28 and S29), demonstrating that the TTA-UC processes of A-1 and A-2 are more efficient than that of DPA.Indeed, the UC emissions of both A-1 and A-2 when S-2 was the sensitizer were much higher than that of DPA (Figure 4a), which was as expected, as the TTET process should be significantly facilitated by the host-guest interactions in A-1 and A-2 [46].To prove this, the TTET processes were investigated based on the variation in phosphorescence intensity of S-2 with changes in the concentration of the annihilators.Increasing the concentration of A-1, A-2, and DPA led to a progressive decrease in phosphorescence intensity, and the Stern-Volmer analysis of the phosphorescence quenching data showed an approximately linear relationship (Figure 4b).The KSV values, when A-1 and A-2 were the acceptors, were indeed much higher than that of DPA, demonstrating that the TTET processes were facilitated by the host-guest interactions, which led to higher UC emissions.
Interestingly, the UC emission of A-2 was much higher than that of A-1, e.g., at the concentration of 25 μM, the UC intensity of A-2 was almost doubled when compared with that of A-1 (Figure 4a).We investigated the TTA-UC quantum yield (ΦUC) at different concentrations of annihilators ranging from 10 μM to 80 μM in toluene and found that the ΦUC significantly increased upon adding different concentrations of annihilators.ΦUC reached to Indeed, the UC emissions of both A-1 and A-2 when S-2 was the sensitizer were much higher than that of DPA (Figure 4a), which was as expected, as the TTET process should be significantly facilitated by the host-guest interactions in A-1 and A-2 [46].To prove this, the TTET processes were investigated based on the variation in phosphorescence intensity of S-2 with changes in the concentration of the annihilators.Increasing the concentration of A-1, A-2, and DPA led to a progressive decrease in phosphorescence intensity, and the Stern-Volmer analysis of the phosphorescence quenching data showed an approximately linear relationship (Figure 4b).The K SV values, when A-1 and A-2 were the acceptors, were indeed much higher than that of DPA, demonstrating that the TTET processes were facilitated by the host-guest interactions, which led to higher UC emissions.M −1 s −1 for A-2) due to the slightly larger binding constants, demonstrating that the TTE efficiency is higher for A-1 than for A-2.But, for TTA-UC emissions, A-1 was much lowe We chose the other complex S-1 with no supramolecular binding site as the sensitizer an found that, by adding the same concentrations of A-1 and A-2, the residual phosphorescen of S-2 was almost the same, demonstrating a comparable TTET process.However, the U emission of A-2 was also much higher than A-1 (Figure 4d).The efficiency of TTA-UC fluorescence is an accumulative result of the efficiencies all of the processes involved in the upconversion, and the ΦUC was generally written a ΦUC = ΦISC × ΦTTET × ΦTTA × ΦF, where ΦISC, ΦTTET, ΦTTA, and ΦF are the quantum yields intersystem crossing of the sensitizer, triplet energy transfer from the sensitizer to the an nihilator, the acceptor triplet annihilation, and acceptor fluorescence, respectively.Con sidering the even lower TTET process for A-2 than A-1 when S-2 was the sensitizer an the comparable fluorescence quantum yield of A-1 and A-2 (according to the above equ tion), the much higher ΦUC of A-2 was, therefore, ascribed to the significantly improve Interestingly, the UC emission of A-2 was much higher than that of A-1, e.g., at the concentration of 25 µM, the UC intensity of A-2 was almost doubled when compared with that of A-1 (Figure 4a).We investigated the TTA-UC quantum yield (Φ UC ) at different concentrations of annihilators ranging from 10 µM to 80 µM in toluene and found that the Φ UC significantly increased upon adding different concentrations of annihilators.Φ UC reached to a plateau at the concentration of 80 µM for both A-1 and A-2 (Figure 4c).It was interesting to find that, at any concentration, the Φ UC of A-2 was constantly higher than that of A-1, and the maximum Φ UC was determined to be 13.7% for A-2, which was more than two-fold higher than that of A-1 (5.1%).These results were a little bit unexpected, as, for both A-1 and A-2, host-guest interaction with the sensitizer S-2 existed and binding constants were comparable.A-1 showed even higher K SV than A-2 (K SV = 1.4 × 10 4 M −1 s −1 for A-1 and 8.5 × 10 3 M −1 s −1 for A-2) due to the slightly larger binding constants, demonstrating that the TTET efficiency is higher for A-1 than for A-2.But, for TTA-UC emissions, A-1 was much lower.We chose the other complex S-1 with no supramolecular binding site as the sensitizer and found that, by adding the same concentrations of A-1 and A-2, the residual phosphorescence of S-2 was almost the same, demonstrating a comparable TTET process.However, the UC emission of A-2 was also much higher than A-1 (Figure 4d).
The efficiency of TTA-UC fluorescence is an accumulative result of the efficiencies of all of the processes involved in the upconversion, and the Φ UC was generally written as Φ UC = Φ ISC × Φ TTET × Φ TTA × Φ F , where Φ ISC , Φ TTET , Φ TTA , and Φ F are the quantum yields of intersystem crossing of the sensitizer, triplet energy transfer from the sensitizer to the annihilator, the acceptor triplet annihilation, and acceptor fluorescence, respectively.Considering the even lower TTET process for A-2 than A-1 when S-2 was the sensitizer and the comparable fluorescence quantum yield of A-1 and A-2 (according to the above equation), the much higher Φ UC of A-2 was, therefore, ascribed to the significantly improved Φ TTA .We carefully quantified the efficiencies of the main processes (TTET and TTA processes) involved in TTA-UC, as shown in Figure 5, it was found that, when S-2 was the sensitizer, the TTET efficiency of A-1 (80 µM) was determined to be 52.9%,which was higher than that of A-2 (38.7%).However, for the TTA efficiencies, only 13.4% was observed for A-1, but 53.4% was observed for A-2; the TTA process in A-2 was much more efficient than that of A-1.Similar results were observed when S-1 was the sensitizer (Figures S30 and S34): the TTA efficiency of A-2 was also three-fold higher than that of A-1 (16.3% for A-1 and 51.4% for A-2).As mentioned earlier, for the TTA process, the following aspects should jointly contribute to the efficiency: the competition between annihilation and decay for the fate of triplet states, the collision probability of the two triplet states of the annihilator, and the spin statistical factor (η) [38].Herein, the first influencing factor should be the same, as same fluorophore DPA was used in both A-1 and A-2.However, the orientations of the two DPA units in A-1 and A-2 were different: the orientation of the two DPA units in A-1 was random, while, in A-2, the two DPA units were pushed to a parallel arrangement by intramolecular hydrogen-bonding interactions.The collision probability of the triplet states of DPA units in A-2 should be higher due to the close proximity of the two DPA units.Also, the probability of forming a singlet state from a pair of annihilating triplet states (the spin statistical factor η) should also be higher, according to a recent theoretical simulation that showed that parallel chromophores gave higher η than perpendicular chromophores [38].Thus, for the first time, we experimentally confirmed that the parallel orientation of the two annihilators significantly enhanced the TTA efficiencies.ΦTTA.We carefully quantified the efficiencies of the main processes (TTET and TTA processes) involved in TTA-UC, as shown in Figure 5, it was found that, when S-2 was the sensitizer, the TTET efficiency of A-1 (80 μM) was determined to be 52.9%,which was higher than that of A-2 (38.7%).However, for the TTA efficiencies, only 13.4% was observed for A-1, but 53.4% was observed for A-2; the TTA process in A-2 was much more efficient than that of A-1.Similar results were observed when S-1 was the sensitizer (Figures S30 and S34): the TTA efficiency of A-2 was also three-fold higher than that of A-1 (16.3% for A-1 and 51.4% for A-2).As mentioned earlier, for the TTA process, the following aspects should jointly contribute to the efficiency: the competition between annihilation and decay for the fate of triplet states, the collision probability of the two triplet states of the annihilator, and the spin statistical factor (η) [38].Herein, the first influencing factor should be the same, as same fluorophore DPA was used in both A-1 and A-2.However, the orientations of the two DPA units in A-1 and A-2 were different: the orientation of the two DPA units in A-1 was random, while, in A-2, the two DPA units were pushed to a parallel arrangement by intramolecular hydrogen-bonding interactions.The collision probability of the triplet states of DPA units in A-2 should be higher due to the close proximity of the two DPA units.Also, the probability of forming a singlet state from a pair of annihilating triplet states (the spin statistical factor η) should also be higher, according to a recent theoretical simulation that showed that parallel chromophores gave higher η than perpendicular chromophores [38].Thus, for the first time, we experimentally confirmed that the parallel orientation of the two annihilators significantly enhanced the TTA efficiencies.To further confirm the pivotal effect of the parallel arrangement of two DPA units by hydrogen-bonding interaction on the TTA efficiencies, the solvent for TTA-UC was changed to MeOH, which is well-known to destroy the hydrogen-bonding interactions.It was found that the UC emissions of both A-1 and A-2 in MeOH were significantly decreased, but the intensities for the two compounds were comparable (Figure S35).More interestingly, we gradually added trace methanol, from 0.33% to 2.00%, to the toluene solution of A-2/S-2, i.e., concentrations for which we thought the polarity of the solvent would not change much.The UC emission of A-2 was significantly decreased (Figure 6a), while adding the same amount MeOH to the A-1/S-2 system did not result in such change-the UC emission basically remained unchanged (Figure 6b).These results demonstrated again that the more efficient TTA-UC for A-2 was due to parallel arrangement of DPA units by hydrogen-bonding interactions.To further confirm the pivotal effect of the parallel arrangement of two DPA units by hydrogen-bonding interaction on the TTA efficiencies, the solvent for TTA-UC was changed to MeOH, which is well-known to destroy the hydrogen-bonding interactions.It was found that the UC emissions of both A-1 and A-2 in MeOH were significantly decreased, but the intensities for the two compounds were comparable (Figure S35).More interestingly, we gradually added trace methanol, from 0.33% to 2.00%, to the toluene solution of A-2/S-2, i.e., concentrations for which we thought the polarity of the solvent would not change much.The UC emission of A-2 was significantly decreased (Figure 6a), while adding the same amount MeOH to the A-1/S-2 system did not result in such change-the UC emission basically remained unchanged (Figure 6b).These results demonstrated again that the more efficient TTA-UC for A-2 was due to parallel arrangement of DPA units by hydrogen-bonding interactions.

Materials and Methods
All chemicals were obtained from commercial suppliers and used as received. 1HNMR and 13 C NMR spectra in CDCl3, DMSO were obtained using a Brucker DRX-400 spectrometer (Billerica, MA, USA).HRMS was determined by MALDI-TOF-MS.UV-vis spectra and fluorescence spectra were recorded using a Duetta spectrometer (Horiba, Kyoto, Japan).Upconversion measurements were performed using a diode-pumped solid-state laser (589 nm or 532 nm) (Minghui Optoelectronic Technology, Xi'an, China), synchronized with a fluoromax-4 fluorescence spectrometer (Horiba, Kyoto, Japan).Fluorescence and upconversion lifetime attenuation are also performed on fluoromax-4 fluorescence spectrometers.

Synthesis of A-1 and A-2
The syntheses of intermediates M1, M2, and P5-Q are shown in the support information.

Materials and Methods
All chemicals were obtained from commercial suppliers and used as received. 1HNMR and 13 C NMR spectra in CDCl 3 , DMSO were obtained using a Brucker DRX-400 spectrometer (Billerica, MA, USA).HRMS was determined by MALDI-TOF-MS.UV-vis spectra and fluorescence spectra were recorded using a Duetta spectrometer (Horiba, Kyoto, Japan).Upconversion measurements were performed using a diode-pumped solid-state laser (589 nm or 532 nm) (Minghui Optoelectronic Technology, Xi'an, China), synchronized with a fluoromax-4 fluorescence spectrometer (Horiba, Kyoto, Japan).Fluorescence and upconversion lifetime attenuation are also performed on fluoromax-4 fluorescence spectrometers.

Measurement
The fluorescence quantum efficiency (Φ F ) of the annihilators and sensitizers was determined by using 9,10-diphenyl anthracene (DPA) in ethanol (Φ F = 95%) and rose bengal (Φ F = 11%) as the standard, respectively.The upconversion quantum yields (Φ UC , of 100% maximum) were determined with cresol violet (Φ std = 51% in EtOH) as the standard and were calculated with the following formula, Equation (1), where Φ UC stands for the upconversion quantum yield; A, I and η represent the absorbance, integrated photoluminescence intensity, and refractive of the solvent, respectively; sam means the sample; and std is the standard.
The efficiencies of the triplet-triplet energy transfer (TTET) between the sensitizers and annihilators were determined by the quenching of the phosphorescence of the sensitizers and was calculated with the formula in Equation ( 2), where I and I 0 stand for the integration of the emission in the presence and absence of the annihilators, respectively.
The TTA efficiencies between the annihilators were determined using Equation (3), where Φ F is the fluorescent quantum yield of the annihilator and Φ ISC is the intersystem crossing (ISC) efficiency of the sensitizer, which was approximately 1.

Conclusions
In conclusion, two supramolecular annihilators, A-1 and A-2, with two DPA units grafted on pillar [5]arenes were synthesized.In A-1, the orientation of the two DPA units was random, while, in A-2, the two DPA units were pushed to a parallel arrangement by intramolecular hydrogen-bonding interactions in toluene.It was found that A-1 and A-2 showed very similar photophysical properties, including similar fluorescence quantum yield and lifetimes.Furthermore, they also showed comparable host-guest binding affinities toward an alkyl ammonia chain-attaching sensitizer S-2.The supramolecular annihilators A-1 and A-2 showed much more efficient TTA-UC compared with the traditional DPA annihilator.The host-guest interaction between the sensitizer and annihilator significantly facilitated the TTET process.More significantly, the TTA-UC emission of A-2 was much higher than that of A-1, and the UC quantum yield of A-2 sensitized by S-2 was determined to be 13.7%, which is more than twice that of A-1 (5.1%).The parallel arrangement of the two DPA units by multiple hydrogen-bonding interactions, which significantly increased the TTA efficiency, was demonstrated to be responsible for the much higher TTA-UC emissions for A-2.This work presented a novel strategy for enhancing TTA-UC efficiency and provided design rationales for innovative supramolecular annihilator.

Table 1 .
Photophysical parameters of the annihilators and the sensitizers a .

Table 1 .
Photophysical parameters of the annihilators and the sensitizers a .
a 1 e The lifetime.