Boosting the circularly polarized luminescence of pyrene-tiaraed pillararenes through mechanically locking

Abstract

Attributed to their unique dynamic planar chirality, pillar[n]arenes, particularly pillar[5]arenes, have evolved as promising platforms for diverse applications such as circularly polarized luminescence (CPL) emitters. However, due to the unit flipping and swing, the achievement of excellent CPL performances of pillar[5]arenes in solution state remains a formidable challenge. To deal with this key issue, a mechanically locking approach has been successfully developed, leading to boosted dissymmetry factor (g_lum) values of pyrene-tiaraed pillar[5]arenes up to 0.015 through the formation of corresponding [2]rotaxanes. More importantly, taking advantage of the stably locked co-conformers, these resultant [2]rotaxanes maintain excellent CPL performances in diverse solvents and wide range of concentrations, making them promising candidates for practical applications. According to this proof-of-concept study, we have not only successfully developed a powerful strategy for the rational design of chiral luminescent materials with desired CPL performances but also contributed a promising platform for the construction of smart chiral materials.

Due to their dynamic planar chirality, pillar[n]arenes are promising circularly polarized luminescence emitters but efficient solution state emission is a challenge due to the unit flipping and swing. Here, the authors report a mechanically locking approach for boosted dissymmetry factor values of pyrene-tiaraed pillar[5]arenes up to 0.015 through the formation of corresponding [2]rotaxanes.

Introduction

Attributed to their potential applications in diverse fields such as advanced display technologies, information encryption, chiral sensing, etc., recently circularly polarized luminescence (CPL) materials have attracted significant attention^1–4. During the past few decades, aiming at realizing both high photoluminescence quantum yields (PLQYs) and high dissymmetry factors (g_lum) for practical applications, diverse CPL materials, such as lanthanide complexes^5–8, small organic^9–12 and organometallic molecules^13–15, polymers^16–19, and aggregates^20–24 have been successfully developed. Notably, along with the blooming of CPL research, the rapid development of supramolecular and macrocyclic chemistry continually contributes to new design strategy and material basis^25–29. In particular, due to their well-defined chiral nanospaces, attractive host-guest properties, and unique hierarchical self-assembly capabilities, chiral macrocycles have proven to be promising platforms for the development of novel CPL materials at different scales^30–36.

Among diverse chiral macrocycles, since the pioneering work by Chen et al.³⁷, pillararene-based CPL materials have attracted more and more attentions due to not only pillararenes feature unique dynamic planar chirality but also they are easy to be prepared, resolved, and functionalized^38–42. However, for chiral pillararenes, conformation flexibility such as unit flipping and swing lead to weaken effects in their CPL performances. Although the unit flipping can be successfully inhibited by introducing bulky substituent groups, the swing of the units remains⁴³. That is why for pillararene-based CPL materials, most of their g_lum values measured in solution state are less than 8.0 × 10^{−3 44–51}. Thus, it is urgent to develop a new strategy for improving the CPL performances of chiral pillararenes, particularly in solution state, which will be of great importance for the further extension of their potential applications.

Based on our ongoing research interest in mechanically interlocked molecules (MIMs), especially rotaxane dendrimers^52–60, previously we successfully realized the construction of a pillararene-based CPL switch through the precise modulation of chirality information transfer from chiral pillararene wheel to the AIEgen core within a AIE-active [3]rotaxane, in which high g_lum value up to 0.014 was successfully achieved in casted films⁶¹. Herein, on the basis of this early exploration, we propose a mechanically locking approach for the construction of pillararene-based CPL materials. In our design strategy, the formation of [2]rotaxanes through mechanical bond would efficiently inhibit the units swing and lock the conformations of chromophore-functionalized pillar[5]arene, thus leading to boosted CPL performances with high g_lum values and PLQYs. To our delight, compared with the chirality-inducing and configuration-fixing strategy, such a design strategy has been proven be feasible and powerful, enabling the successful construction of pillar[5]arene-based [2]rotaxanes with excellent CPL performances with g_lum values up to 0.015 in solution state, making them attractive platforms for practical use (Fig. 1).

Fig. 1. Design concept of this work.

Design strategy for boosting the CPL performances of chromophore-functionalized pillar[5]arenes through mechanically locking approach, as compared with chirality inducing and configuration fixing strategy.

Results

Design, synthesis, and characterization of penta(pyrenyl)pillar[5]arenes and corresponding [2]rotaxanes

In our study, pyrene unit was chosen as the emission moiety since pyrene excimers have been proven to be excellent CPL emitters with promising performances^62–68. Starting from a rim-differentiated tiara-pillar[5]arene Br-P5-OMe with five 2-bromoethoxy moieties^69,70, the corresponding penta(pyrenyl)pillar[5]arene⁷¹ Py-P5-OMe was successfully synthesized in 79.0% yield on a gram scale through etherification reaction with 1-pyrenol (Fig. 2 and Supplementary Fig. S1), which serves as not only the wheel component for the synthesis of corresponding [2]rotaxanes but also a control compound for the chirality inducing approach⁷². In addition, to further evaluate the effect of unit flipping on the CPL performances, cyclohexylmethoxy units were then introduced as bulky substituent groups for the synthesis of pyrene-tiaraed pillar[5]arene. Thus, a new rim-differentiated tiara-pillar[5]arene Br-P5-Cy was then synthesized, which was confirmed by its single-crystal X-ray structures (Supplementary Fig. S102 and Table S12). Under similar condition with that of Py-P5-OMe, the corresponding penta(pyrenyl)pillar[5]arene Py-P5-Cy was prepared in 59.0% yield and confirmed by single-crystal X-ray study (Supplementary Fig. S103 and Table S12) as a control compound for configuration fixing approach (Fig. 2 and Supplementary Fig. S2)⁷³. Both penta(pyrenyl)pillar[5]arenes have been well characterized by various techniques including 1-D multinuclear (¹H and ¹³C), 2-D NMR (¹H-¹H COSY, NOESY) spectroscopy, and HR ESI-MS analysis (Supplementary Fig. S4-S23).

Fig. 2. Synthesis of penta(pyrenyl)pillar[5]arenes and corresponding [2]rotaxanes.

Synthesis of penta(pyrenyl)pillar[5]arenes Py-P5-OMe, Py-P5-Cy (with its single-crystal X-ray structure), and [2]rotaxanes Py-[2]R-10C, Py-[2]R-8C. Reaction conditions: a 1-pyrenol, K₂CO₃, DMF, 85 °C, 48 h, 79.0%; b 1-pyrenol, K₂CO₃, DMF, 85 °C, 48 h, 59.0%; c sebacoyl dichloride (for Py-[2]R-10C) or suberoyl dichloride (for Py-[2]R-8C), pentafluorophenol, −45 °C to r.t., Et₃N, CHCl₃, overnight, 15.5% for Py-[2]R-10C and 9.5% for Py-[2]R-8C.

To ensure the locking of stable chiral conformers of penta(pyrenyl)pillar[5]arene, we envisaged that the formation of rotaxanes could efficiently inhibit the unit flipping and significantly reduce the degree of unit swing of penta(pyrenyl)pillar[5]arene through the existence of axle component within the cavity, thus leading to stronger CPL emission. To evaluate the feasibility of such mechanically locking approach, the synthesis of penta(pyrenyl)pillar[5]arene-based rotaxanes was then performed. In addition, to further reduce the flexibility of the pyrene units and promote their chiral stacking, pentafluorobenzene unit was chosen as the stopper^74–76. On the one hand, pentafluorobenzene is big enough for preventing the de-threading of axle and small enough for not disturbing the formation of pyrene excimers. On the other hand, pentafluorobenzene could help to stabilize and lock the chiral conformations through the possible formation of a sandwich-like D-A-D stacking with two pyrene units.

According to such design strategy, to check the feasibility of the formation of mechanical bond⁷⁷, the synthesis of the model [2]rotaxane Br-[2]R-10C with Br-P5-OMe as the wheel component was first carried out (Supplementary Fig. S3). To our delight, the reaction of sebacoyl chloride with pentafluorophenol in the presence of 2 eq. Br-P5-OMe and 2.5 eq. Et₃N in CHCl₃ gave rise to the desired [2]rotaxane Br-[2]R-10C in 18.3% yield, and its chemical structure was unambiguously confirmed by the single-crystal X-ray analysis (Supplementary Fig. S105 and Table S13). On the basis of such synthetic exploration, the employment of Py-P5-OMe as the wheel component under the same condition resulted in the successful corresponding [2]rotaxane Py-[2]R-10C with a slight decrease in the yield (15.5%) due to the larger steric hindrance of five attached pyrene units. In addition, to investigate the length effect of the axle component on the conformation locking, sebacoyl dichloride was then replaced by suberoyl dichloride for the [2]rotaxane synthesis, leading to the synthesis of Py-[2]R-8C in expected even reduced yield of 9.5% (Fig. 2 and Supplementary Fig. S3)^78,79. As expected, due to the even larger steric hindrance in both rims of Py-P5-Cy, the synthesis of corresponding [2]rotaxanes failed even after continuous attempts.

For both [2]rotaxanes Py-[2]R-10C and Py-[2]R-8C, characteristic peaks below 0.0 ppm, which are typical signals for pillar[5]arene/alkyl chain rotaxanes, were observed in their ¹H NMR spectra (Supplementary Figs. S33 and S42), indicating the successful formation of the interlocked structures. In addition, compared with the individual thread and macrocycle components, remarkable changes in chemical shifts, particularly downfield shifts of the signals attributed to pillar[5]arene units and upfield shifts of the signals ascribed to the alkyl chain units (Supplementary Figs. S36 and S45), were observed, again suggesting the formation of targeted [2]rotaxanes. In the 2D ¹H-¹H NOESY spectra, correlation signals between the methylene protons and protons of Py-P5-OMe (for Py-[2]R-10C, H_e and H_A, H_b and H_C, H_f and H_B) (Supplementary Fig. S41) were found, again indicating the existence of mechanical bond between the axle and wheel components. More importantly, high-resolution ESI-TOF-MS analysis further confirmed the successful synthesis of these [2]rotaxanes (Supplementary Figs. S29, S39, and S47). For instance, for Py-[2]R-10C, the main characteristic peak at m/z = 2457.7922 was observed, which was attributed to the [Py-[2]R-10C + Na]⁺ ion peak and fully consistent with the theoretical value (m/z = 2457.7849).

After successful synthesis of the targeted [2]rotaxanes, to investigate the expected mechanical bond-induced conformation locking within their interlocked structures, single crystal growth for X-ray diffraction analysis was then performed. After continuous attempts, single crystal growth of Py-[2]R-8C succeed by slow diffusion of n-hexane into its dichloromethane (DCM) solution (Supplementary Figs. S106–S108 and Table S13). While in the case of Py-[2]R-10C, possibly due to the longer axle component that makes the dense packing become more difficult, no suitable single crystals were obtained. Similar with the case of Py-P5-Cy, for the single crystal structure of Py-[2]R-8C obtained from the racemic mixture, a pair of enantiomers with the Py-P5-OMe wheels in either (pS, pS, pS, pS, pS) or (pR, pR, pR, pR, pR) conformers holding together by multiple C-H···π interactions were observed (Fig. 3a), which is reasonable since these two conformers are in lower energy and with larger cavity compare with other ones. While in Py-P5-Cy, an intermolecular pyrene dimer with a distance of 3.554 Å serves as the “bridge” to stabilize the two enantiomer molecules (Supplementary Fig. S104). Interestingly, in the [2]rotaxane molecule, the expected sandwich-like stacking of pentafluoro-benzene and pyrene units at positions 1 and 3 with the distances of 3.454 Å and 3.633 Å was observed (Fig. 3b), which is in consistent with our design strategy and further supported by ¹⁹F NMR spectra comparison of [2]rotaxanes Br-[2]R-10C and Py-[2]R-10C (Supplementary Fig. S37). In addition, to form a dense packing in the signal crystal, pyrene units at positions 4 and 5 serve as a “tweezer” to give rise to an intermolecular D-A-D stacking with a free pentafluorobenzene stopper from neighboring [2]rotaxane molecule with the distances of 3.480 Å and 3.523 Å (Fig. 3c). In the case of pyrene unit in position 2, an intermolecular dimer with a distance of 3.480 Å but very little overlap is formed by two neighboring [2]rotaxane molecules (Fig. 3d, e). According to these packing manners, for the single crystal, no pyrene excimer emission was observed. However, in solution state, the intermolecular stackings would be weakened while intramolecular ones would be strengthened, thus leading to expected locked pyrene excimer stacking with strong CPL emission.

Fig. 3. Single-crystal X-ray structure of Py-[2]R-8C.

Capped sticks representation model of the single-crystal X-ray structure, intra- and inter-molecular stacking of pyrene units of Py-[2]R-8C. a A pair of enantiomers of Py-[2]R-8C (red: pR-Py-[2]R-8C; blue: pS-Py-[2]R-8C) in a single cell axe; b The intramolecular stacking between pentafluorophenol (yellow) and pyrene 1 and 3 (green); c The intermolecular stacking between pentafluorophenol (purple) and pyrene 4 and 5 (cyan); d The intermolecular stacking between pyrene 2 (pink) from two different [2]rotaxane molecules with different planar chirality; e The misaligned π–π stacking between pyrene 2 (pink) from two different [2]rotaxane molecules with different planar chirality.

The CPL performances of penta(pyrenyl)pillar[5]arenes as control compounds under chirality inducing and configuration fixing

For Py-P5-OMe, five pyrene units have been introduced, thus providing an attractive platform for the construction of CPL-active materials through the formation of well-defied chiral pyrene excimer stacking. However, since the other rim of Py-P5-OMe is methoxy group, which is too small to lock the chiral configuration, Py-P5-OMe is CPL-silent. According to previous reports⁷², the addition of special chiral guests might induce the formation of chiral conformers that further give rise to chiroptical response such as CPL emission, chiral amino acid derivatives D-Ala-OEt were employed as the guest molecule for the complexation study. Unfortunately, although the successful formation of 1:1 host-guest complex between Py-P5-OMe and D-Ala-OEt was suggested and the binding constant was determined to be 1.19 × 10⁴M⁻¹ (Supplementary Figs. S50–S53), Circular Dichroism (CD) and Circularly Polarized Luminescence (CPL) spectra revealed no induced chiroptical signals even after the addition of much excess chiral guests (Supplementary Figs. S54–S56). These results indicated the chirality-inducing approach failed, which might be attributed to the lack of additional noncovalent recognition sites that could promote the formation of stable chiral conformations. In the resultant host-guest complex, the unit flipping and swing of Py-P5-OMe remain, resulting in no obvious chiral induction for CPL emission.

In the case of Py-P5-Cy, the existence of bulky cyclohexylmethoxy units enables the successful chiral resolution with a semi-preparative CHIRALPAK IA® column, and two fractions with the e.e. values > 99% were isolated (Supplementary Figs. S57–S60). According to the solution-state CD spectra, mainly the bands at both 250 nm and 310 nm attributed to the pillar[5]arene skeleton, the absolute configurations of enantiomers in first fraction and second fraction were assigned to be pR and pS (Supplementary Fig. S61), respectively^80–83. More importantly, in the CD spectra, several major bands in the 320–400 nm range were found, which were ascribed to the pyrene units as the typical vibronic π–π* bands, thus clearly indicating the chirality transfers from the chiral pillar[5]arene skeleton to the attached pyrene units in the ground state. In addition, the 0–0 UV bands fit very well with the corresponding first Cotton CD bands at 383 nm, indicating the expected chiral stackings of pyrene excimers induced by the chiral pillar[5]arene skeleton with fixed configurations (Fig. 4a, left).

Fig. 4. Chiroptical properties of penta(pyrenyl)pillar[5]arene and [2]rotaxanes.

CD/UV (left), CPL/FL (middle), and g_lum spectra of Py-P5-Cy (a), Py-[2]R-10C (b), and Py-[2]R-8C (c) in CH₂Cl₂ (0.01 mM, 298 K) upon excitation at 350 nm.

Interestingly, The CPL tests of pR/pS-Py-P5-Cy (DCM, 10 μM) revealed mirrored-imaged bisignate CPL peaks with the crossover at 464 nm (Fig. 4a, middle), again indicating the successful isolation of a pair of enantiomers. Notably, due to the existence of five pyrene units, the interpretation of the observed couplet in CPL spectra is difficult. Since the emissions at both 433 nm and 524 nm are attributed to the pyrene excimers according to the emission lifetime measurements (15.6 ns at 433 nm and 22.3 ns at 524 nm, Supplementary Fig. S65, Table S1), the possible explanation might be the co-existence of two pairs of excimers in the excited states. The |g_lum| values at 433 nm and 524 nm were measured to be 3.8 × 10⁻³ and 2.2 × 10⁻³ (Fig. 4a, right), respectively, which are comparable with previous study of planar chiral pillar[5]arenes. The further increasing (to 100 μM) or reducing (to 1 μM) the concentration did not lead to significant changes in neither g_lum values nor PLQYs (all around 35%) (Supplementary Fig. S64). According to these results, compared with the chirality-inducing approach, the fixed configurations did lead to the emergence of CPL emissions by inhibiting the unit flipping. However, the remained unit swing resulted in moderate CPL performances among diverse chiral pillar[5]arenes.

Boosted CPL performances of penta(pyrenyl)pillar[5]arene-based [2]rotaxanes through mechanically locking

To further evaluate the conformation-locking effect on the CPL performances in solution state, the optical resolution of [2]rotaxanes Py-[2]R-10C and Py-[2]R-8C was then carried out. For both [2]rotaxanes, by eluting with EtOAc/n-hexane (25/75, V/V) with a semi-preparative CHIRALPAK IF® column, two fractions were successfully isolated with the e.e. values of each fraction all above 99%, indicating the successfully obtainment of a pair of optical pure [2]rotaxanes (Supplementary Figs. S66–S73), which is in consistent with single crystal X-ray analysis. Similar with Py-P5-Cy, the absolute configurations of the resultant chiral [2]rotaxanes were then assigned to be pR and pS for the one in first fraction and second fraction, respectively (Supplementary Figs. S74, S75). As revealed by CD spectra (Fig. 4b, c, left), by normalizing the intensity of bands at 310 nm for pillar[5]arene skeleton, compared with Py-P5-Cy, higher intensities of major bands in the 320-400 nm range attributed to the pyrene units were observed for both chiral [2]rotaxanes. In particular, the dissymmetry factors |g_abs| at 383 nm of Py-P5-Cy, Py-[2]R-10C, and Py-[2]R-8C were calculated to be 1.04 × 10⁻⁴, 1.09 × 10⁻⁴, and 1.27 × 10⁻⁴, respectively, indicating gradually enhanced Cotton effects. According to these results, the formation of chiral [2]rotaxanes was proven successfully induce the well-defied chiral pyrene stackings with enhanced chirality information transfer, preliminarily suggesting the crucial role of mechanical bond for conformation locking.

To our delight, in the CPL spectra (DCM, 10 μM), intense peaks at 485 nm attributed to pyrene excimer emission were observed for both Py-[2]R-10C and Py-[2]R-8C in a mirror-imaged form (Fig. 4b, c, middle), suggesting the excited state chirality transfer from chiral pillar[5]arene to pyrene units. More importantly, the g_lum values of Py-[2]R-10C were up to 0.013 at 455 nm, and for Py-[2]R-8C, the values were even larger (up to 0.015 at 455 nm) (Fig. 4b, c, right). These values are not only quite large compared with traditional small organic molecules with g_lum values typically on the order of 10⁻⁵ to 10⁻³, but also among the largest one for reported pillararene-based CPL systems in solution state (Table S11). Moreover, the CPL brightness^84–86 B_CPL of Py-[2]R-10C and Py-[2]R-8C were calculated to be 263.0 M⁻¹ cm⁻¹ and 364.2 M⁻¹ cm⁻¹, respectively, and such high values again highlight the excellent CPL performances of these chiral [2]rotaxanes.

Notably, considering the possible influence of solvent on the CPL performances, solvent-dependent CPL measurements were then performed (Supplementary Figs. S92–S95 and Tables S6–S8). For both [2]rotaxanes, in all ten tested solvents (DCM, EA, THF, toluene, 1,4-dioxane, DMF, DMSO, n-hexane, MeCN, and acetone), the g_lum values were all above 8.0 × 10⁻³. In addition, in seven out of ten tested solvents, the g_lum values were more than 0.01, highlighting the outstanding CPL performances of these chiral [2]rotaxanes (Supplementary Fig. S96 and Table S8). Notably, since under the CPL test condition with long-time 350 nm UV light irradiation, chloroform might produce acidic species that would lead to the degradation of [2]rotaxanes, CPL tests in chloroform did not give rise to satisfactory results although it is the most commonly-used solvent for pillararene study (Supplementary Figs. S97, 98). Attractively, for these [2]rotaxanes, their PLQYs are relatively high (e.g., in DCM: 35.1% for Py-[2]R-10C and 38.3% for Py-[2]R-8C; in DMSO: 44.3% for Py-[2]R-10C and 42.2% for Py-[2]R-8C), which are almost identical with that of the Py-P5-OMe wheel (33.4%), suggesting the maintained fluorescence emission efficiency along with the formation of [2]rotaxanes (Supplementary Table S9). The emission lifetime measurements indicated that, for both [2]rotaxanes, the emission lifetimes at 485 nm and 455 nm were all above 20 ns (Py-[2]R-10C, 23.0 ns at 455 nm and 22.9 ns at 485 nm; Py-[2]R-8C, 22.6 ns at 455 nm and 22.0 ns at 485 nm), confirming these emissions are all originated from pyrene excimers (Supplementary Figs. S99–S101 and Table S10). In addition, the emission lifetimes at 400 nm were measured to be 3.30 ns for Py-[2]R-10C and 4.48 ns for Py-[2]R-8C, both of which could be assigned as monomer emissions. Notably, although a lot of pillararene-based rotaxanes have been reported, but their CPL performances have been rarely explored, the excellent CPL performances of these chiral [2]rotaxanes are quite attractive.

According to these promising results, our proposed mechanically locking approach has been proven to be effective for the achievement of both high PLQY and g_lum. Upon the formation of mechanical bond, different with the stacking in the crystal, attributed to the strong π–π stacking, pyrene units at positions 4 and 5 would give rise to stable chiral excimer stacking with inhibited unit flipping and negligible unit swing due to the locking effect of the mechanical bond, thus resulting in excellent CPL performances in solution state. In addition, compared with Py-[2]R-10C, the shorter axle in Py-[2]R-8C might further promote this chiral excimer stacking through more stable co-conformer locking, thus leading to even better CPL performances.

To further provide more support for such design strategy, concentration-dependent FL and CPL spectra were then measured (Supplementary Figs. S81–S87). In the FL spectra, to gain more in-depth understanding of the conformations, attention was paid on the intensity of excimer emission at 485 nm and monomer emission at 386 nm. At the concentration of 10 μM, the excimer vs. monomer emission I_e/I_m ratios of Py-[2]R-10C and Py-[2]R-8C were 3.71 and 3.56, respectively. Upon gradually increasing the concentration to 1000 μM, these values increased to 88.15 and 62.48, respectively, which were reasonable due to the formation of diverse intermolecular pyrene excimers in high concentration. As expected, upon diluting the concentration to 0.4 μM, these values slightly decreased to 2.91 and 2.61, respectively (Supplementary Tables S2 and S3). More interestingly, for both [2]rotaxanes, at the concentration range from 0.4 μM to 2.0 μM, the I_e/I_m ratios were almost the same, indicating the existence of stable intramolecular pyrene excimer conformations (blue box in Fig. 5a). Upon increasing the concentration from 4.0 μM to 40 μM, the I_e/I_m ratios are also gradually enhanced, suggesting the emergence of intermolecular pyrene excimers of the pyrene unit at position 2 (red box in Fig. 5a), which is similar with the case in single crystal. Such intermolecular dimer complex remains stable until the concentration reaching to 100 μM, as revealed by the slightly increased I_e/I_m ratios in the range of 40 μM to 100 μM. When the concentrations are above 100 μM, attributed to the formation of complicated intermolecular pyrene excimers (Fig. 5a), the I_e/I_m ratios significantly increased, but with obvious fluorescence quenching. More importantly, the same trend was observed in concentration-dependent CPL spectra. Attributed to stable intramolecular chiral pyrene excimer conformations, as well as the dynamic feature of the intermolecular dimer complex that has slight influence on the CPL performances, the g_lum values of all these chiral [2]rotaxanes remained unchanged at wide concentration range from 0.4 μM to 100 μM (Fig. 5b, c), which is of great importance for practical use. The further increase of the concentration to 1000 μM led to reduced g_lum values, but they are still above 8.0 × 10⁻³ (at 1000 μM, 8.8 × 10⁻³ for Py-[2]R-10C and 9.6 × 10⁻³ for Py-[2]R-8C), which are even larger than most reported ones for planar chiral pillar[5]arenes so far. Notably, in all tested concentrations, the g_lum values of Py-[2]R-8C are larger than that of Py-[2]R-10C, again suggesting the interesting length effect of the axle component on the promotion of conformation locking (Supplementary Table S4).

Fig. 5. Investigations on the mechanically locking approach through concentration-dependent FL and CPL spectra.

a Cartoon representations of the intra- and inter-molecular pyrene excimer stacking of [2]rotaxanes upon the increase of concentration; Concentration-dependent FL spectra (left), g_lum spectra (middle), the concentration-dependent excimer/monomer ratios and g_lum values (right) of Py-[2]R-10C (b), and Py-[2]R-8C (c) in CH₂Cl₂ (298 K) upon excitation at 350 nm.

In addition, temperature-dependent CPL study in MeCN was further carried out. To our delight, as expected, for both [2]rotaxanes Py-[2]R-10C and Py-[2]R-8C, upon gradually cooling to −10 °C, a remarkable increase in |g_lum| values was achieved (for Py-[2]R-10C, from 0.0135 to 0.019; for Py-[2]R-8C, from 0.0144 to 0.020). While upon gradually heating to 70 °C, an obvious decrease in |g_lum| values was recorded (for Py-[2]R-10C, from 0.0135 to 0.0073; for Py-[2]R-8C, from 0.0144 to 0.0078) (Supplementary Fig. S91). These results again supported the mechanical locking effect since temperature did serve as a key factor to influence the stability of the locked co-conformations.

To provide additional insights into such design strategy, computational simulations were then carried out (for details, see Section F in the Supplementary Information). For both Py-[2]R-10C and Py-[2]R-8C, three interconvertible co-conformers were obtained in both the ground state and the excited state. In the ground state, ΔE values are up to 10.65 kcal/mol for Py-[2]R-10C and 10.66 kcal/mol for Py-[2]R-8C (Supplementary Figs. S109, S110), and these values are up to 4.55 kcal/mol for Py-[2]R-10C and 2.24 kcal/mol for Py-[2]R-8C in the excited state (Fig. 6a, d and Supplementary Figs. S111, S112). As expected, for all the simulated co-conformers, different packing models of the pyrene units at positions 4 and 5 were observed, which are mainly responsive for the CPL emissions with varied performances. In particular, for the most stable co-conformers of these two [2]rotaxanes in the excited state (Py-[2]R-10C-TD-III and Py-[2]R-8C-TD-III), very similar pyrene excimers were found (red box in Fig. 6a, d, pink box in Fig. 6b, e). More importantly, regarding the g_lum, the values can be roughly calculated as 4(|μ|·|m|·cosθ)/(|μ|²+|m|²), where m and μ denote the magnetic and electric transition dipole moments, respectively, and θ is the angle between them. For [2]rotaxanes Py-[2]R-10C and Py-[2]R-8C, as listed in Supplementary Tables S14 and S15, their g_lum values were calculated to be 0.014 and 0.0155, which fit very well with the experimental ones (0.013 for Py-[2]R-10C and 0.015 for Py-[2]R-8C).

Fig. 6. Computational simulation supports for the CPL performances of [2]rotaxanes.

Optimized structures with the corresponding energies (using the DFT level with the B3LYP/6-31 g method) of Py-[2]R-10C (a, side view with the pyrene excimer packing being highlighted; b, top view of Py-[2]R-10C-TD-III with the pyrene excimer packing being highlighted; c, side view of Py-[2]R-10C-TD-III with the axle component being highlighted) and Py-[2]R-8C (d, side view with the pyrene excimer packing being highlighted; e, top view of Py-[2]R-8C-TD-III with the pyrene excimer packing being highlighted; f, side view of Py-[2]R-8C-TD-III with the axle component being highlighted) in the excited state. The stacked pyrenes are colored green and orange, and the axle component is colored red. Most hydrogen atoms are omitted for clarity.

In addition, the simulations also provide a nice explanation on the interesting length effect of the axle component. As shown in Fig. 6c, f, for Py-[2]R-10C, the longer axle component (10 methylene units) makes the pentafluorobenzene stopper near the pyrene rim slipping out of the pocket of pyrene units at position 1 and 3. While in the case of Py-[2]R-8C, the shorter axle component (8 methylene units) leads to the formation of D-A stacking between the pentafluorobenzene stopper and the pyrene unit at position 1, thus resulting in better conformation locking effect and thereby giving rising to even better CPL performances.

Discussion

In summary, by developing a mechanically locking approach, we successfully realized the remarkably boosted CPL performances of pyrene-tiaraed pillar[5]arenes in solution state. The introduction of mechanical bond endows the resultant [2]rotaxanes with stably locked co-conformers, thereby leading to enhanced g_lum values of penta(pyrenyl)pillar[5]arenes up to 0.015 through the efficient inhibition of units flipping and swing. Attractively, in addition to relatively high PLQY, these [2]rotaxanes maintain excellent CPL performances in diverse solvents and wide range of concentrations, laying the foundation for further practical uses. Compared with chirality-inducing and configuration-fixing approaches, the mechanically locking approach has proven to be feasible and efficient for the rational design of chiral macrocycles with excellent CPL performances, thus providing not only a feasible and powerful strategy for the construction of chiral luminescent materials but also promising platform for the further explorations of practical applications such as switchable CPL systems. The further extension of this design strategy to other types of MIMs such as molecular links and knots as well as other chiral macrocycles are actively investigated in our laboratory and the results will be reported in due course.

Methods

NMR spectroscopy experiments

¹H NMR and ¹³C NMR spectra were recorded on Bruker 300 MHz Spectrometer (¹H: 300 MHz), Bruker 400 MHz Spectrometer (¹H: 400 MHz; ¹⁹F: 376 MHz) and Bruker 500 MHz Spectrometer (¹H: 500 MHz; ¹³C: 125 MHz) at 298 K. The ¹H, ¹³C, and ¹⁹F NMR chemical shifts are reported relative to residual solvent signals. 2D-NMR spectra (¹H-¹H COSY, NOESY) were recorded on Bruker 500 MHz Spectrometer (¹H: 500 MHz) at 298 K.

Chiral resolution experiments

Chiral resolution was performed using a Shimadzu HPLC System (LC-20AR) equipped with a CHIRALPAK IF® column (250 mm L × 10 mm I.D. for separation and 250 mm L × 4.6 mm I.D. for analysis) and eluting with hexane/dichloromethane at flow rates of 3.0 mL/min for separation and 1.0 mL/min for analysis.

Optical and chiroptical property characterizations

UV–vis spectra and steady-state fluorescence spectra were recorded in a quartz cell (light path 5 mm) on a Shimadzu UV2700 UV–visible spectrophotometer and a Shimadzu RF-6000 fluorescence spectrophotometer. Emission lifetime measurements and fluorescence quantum yields were obtained at ambient temperature in absolutely in CH₂Cl₂ solution using Fluorescence FL-QM-LT Steady state transient fluorescence spectrometer (λ_ex = 375 nm in the test of emission lifetime measurements and λ_ex = 350 nm in the test of fluorescence quantum yields). CD and CPL spectra were measured on a Chirascan Series Spectrometer (Applied Photophysics Ltd, UK) at room temperature (λ_ex = 350 nm). All the CPL data were obtained by scanning three cycles expect for special situations.

Theoretical calculation

All theoretical calculations were performed by the Gaussian 16 package, details of which were given in Supplementary information. The atomic coordinate files obtained after DFT optimization are included as Supplementary Source Data in the Supplementary Files.

Author contributions

W.W., H.-B.Y., and X.-Q.W. conceived and designed the project. J.-L.S. designed and performed the synthetic experiments with the help of C.C. and Z.P., X.L. and Y.J. designed and performed the computational studies. W.W. and J.-L.S. prepared the manuscript with the help of all other authors.

Peer review

Peer review information

Nature Communications thanks Ying Han and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Data availability

The authors declare that the data supporting this study are available within the paper and its supplementary information file. All other data are available from the authors upon request. Materials and methods, experimental procedures, characterization data, NMR spectra, and mass spectrometry data are available in the Supplementary Information. The X-ray crystallographic coordinates for structures reported in this study have been deposited at the Cambridge Crystallographic Data Centre (CCDC), under deposition numbers CCDC 2327874 (Br-P5-Cy), 2341386 (Py-P5-Cy), 2310043 (Br-[2]R-10C), and 2309168 (Py-[2]R-8C). These data can be obtained free of charge from the CCDC via www.ccdc.cam.ac.uk/data_request/cif. Source Data are provided with this manuscript. Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-024-54961-0.