An Organic Molecular Nanobarrel that Hosts and Solubilizes C₆₀

Abstract

Organic cages have gained increasing attention in recent years as molecular hosts and porous materials. Among these, barrel‐shaped cages or molecular nanobarrels are promising systems to encapsulate large hosts as they possess windows of the same size as their internal cavity. However, these systems have received little attention and remain practically unexplored despite their potential. Herein, we report the design and synthesis of a new trigonal prismatic organic nanobarrel with two large triangular windows with a diameter of 12.7 Å optimal for the encapsulation of C₆₀. Remarkably, this organic nanobarrel shows a high affinity for C₆₀ in solvents in which C₆₀ is virtually insoluble, providing stable solutions of C₆₀.

A new trigonal prismatic organic nanobarrel with two large triangular windows and a diameter of 12.7 Å has been synthesized that is optimal for the encapsulation of C₆₀. Remarkably, this organic nanobarrel shows a high affinity for C₆₀ in solvents in which C₆₀ is virtually insoluble, and results in stable solutions of C_60.

Introduction

Subcomponent self‐assembly via dynamic covalent chemistry is a powerful tool to synthesize complex architectures, which otherwise will be difficult to obtain by conventional multistep organic synthesis. ^[1] Among these architectures, organic cages are three‐dimensional molecular structures, defined by a rigid framework made entirely of covalent bonds. ^[2] The structure of the framework generates a well‐defined internal cavity and also windows of different shapes and dimensions that allow entering and exiting the cavity. Owing to these structural features, organic cages present many opportunities as molecular hosts and porous materials, which have a lot of potential in several applications including encapsulation,[ ^2a , ³ ] molecular reaction vessels,[ ^3a , ⁴ ] separations,[ ^3b , ⁵ ] and sensing, ^[6] among others.

The access to the internal cavity of organic cages is limited by the size of the windows. Most organic cages present windows that are smaller than the size of cavity, precluding the diffusion of hosts larger than windows into the cage. Nonetheless, there are also organic cages with windows of the same size as the internal cavity, such as organic nanobarrels, ^[7] which allow the diffusion of larger hosts into the cage. However, despite their potential to encapsulate large molecules and aggregates, as illustrated by recent reports on their coordination homologous, ^[8] organic nanobarrels remain practically unexplored.

The encapsulation of fullerenes has received a lot of attention, ^[9] not only because of the challenges associated to the molecular recognition of their curved aromatic surface, but also, because of the implications that such recognition has in the properties, purification, and processability of fullerenes. While, coordination cages have been widely investigated in the encapsulation of fullerenes,[ ^8c , ^8g , ^9g , ^9h , ¹⁰ ] organic cages are comparatively less explored[ ^7a , ¹¹ ] despite the potential benefits in terms of stability and charge neutrality. This illustrates that the design and synthesis of organic cages with optimal cavities and windows for the encapsulation of fullerenes remains challenging.

Herein, we report the design and synthesis of a new trigonal prismatic organic nanobarrel (1) constituted by three pyrene panels and two large triangular windows with a diameter of 12.7 Å (Figure 1). Nanobarrel 1 has been synthesized by a one‐pot procedure that involves condensation of a newly designed tetratopic pyrene precursor (2) and 1,2‐diaminocyclohexane (3) through imine bonds in a [3+6] stoichiometry without the use of any template. Theoretical calculations show the excellent structural fit between nanobarrel and C₆₀. In fact, complexation studies show the high affinity of nanobarrel 1 for C₆₀ that can be attributed to the optimal diameter of the cage and to the three pyrene walls. Remarkably, the structural complementarity and high affinity allows encapsulating C₆₀ in non‐aromatic solvents, in which C₆₀ is virtually insoluble, providing stable solutions of C₆₀.

Figure 1.

a) General route for the synthesis of nanobarrel 1. b) Side and c) front views of the calculated structure of nanobarrel 1.

Results and Discussion

Precursor 2 was synthesized in three steps (Scheme 1 and S1). First, 4,5,9,10‐tetrabromo‐2,7‐di‐tert‐butylpyrene (4) was synthesized in two steps from pyrene following reported procedures. ^[12] Precursor 2 was obtained by Suzuki coupling between 4,5,9,10‐tetrabromo‐2,7‐di‐tert‐butylpyrene (4) and 4‐formyl phenylboronic acid (5) in a 20 % yield.

Scheme 1.

Synthesis of precursor 2.

The condensation between precursors 2 and 3 in a stoichiometric ratio of 1 : 2, in the presence of a catalytic amount of trifluoroacetic acid (TFA) yields the [3+6] trigonal‐prismatic nanobarrel 1 (Figure 1a). Parameters that could determine the outcome of the reaction include the choice of the solvent, concentration, temperature, and catalyst used. For instance, the wrong solvent choice could lead to premature precipitation of kinetic intermediates from the reaction mixture that may not be able to equilibrate into the desired product. The yield of the reaction largely depends on the choice of the solvent and the reaction conditions (Table S1). We obtain the best yield for nanobarrel 1 (84 %) at room temperature using chloroform as a solvent, by adding dropwise a dilute solution of precursor 3 into a solution of precursor 2 with a catalytic amount of TFA and by stirring the resulting mixture for 5 days.

Remarkably, among the different nanobarrels with varying stoichiometries that could have been formed during the course of the reaction (Figure S1), we have only experimentally observed the corresponding [3+6] trigonal‐prismatic nanobarrel 1. The structure of nanobarrel 1 has been confirmed by ¹H NMR, ¹³C NMR, and matrix‐assisted laser desorption ionization time‐of‐flight mass spectrometry (MALDI‐TOF MS) (Figure 2). The ¹H NMR spectrum of nanobarrel 1 shows the broadening and also the shift and splitting of the aromatic signals when compared to the spectrum of precursor 2 (Figure 2a). These spectral differences are consistent with the formation of a conformationally‐flexible structure. The ¹H NMR spectrum of nanobarrel 1 shows simultaneously a new signal at 8.42 ppm that corresponds to the imine protons (a′) and the disappearance of the signals corresponding to the aldehyde protons (a) of precursor 2 (Figure 2a), confirming the transformation of the aldehydes into imines. This is consistent with the FT‐IR spectrum of nanobarrel 1 shows the imine C=N stretching signal at 1640 cm⁻¹ and the disappearance of the aldehyde C=O and C−H stretching signals of precursor 2 (Figure S2). Furthermore, the ¹H NMR signals of the (4‐formylphenyl)imino substituents (b′, b′′, d′ and d′′) are the result of the splitting of signals of the 4‐formylphenyl substituents of precursor 2 (b and d) that indicates both the formation of the barrel and the proximity to the asymmetric diaminocyclohexane residues (Figure 2a). To further confirm that the broadening of NMR signals is due to conformational flexibility of the barrel and not due to the formation of a complex mixture of nanobarrels with different stoichiometry, a diffusion‐ordered spectroscopy (DOSY) NMR spectrum was recorded. The spectrum shows that all the signals diffuse at the same rate (9.91×10⁻⁷ cm² s⁻¹), which confirms that the signals correspond to a single species of the same stoichiometry (Figure 2b). Variable temperature NMR spectra (293–343 K, tetrachloroethane‐d ₂) illustrate that the broadening of the NMR signals is the result of the conformational flexibility of nanobarrel 1 (Figure 2c and S3). ^[13] The ¹³C NMR spectrum of nanobarrel 1 shows similar features to those observed on the ¹H NMR, including the broadening and shifts of the aromatic signals, and also, the disappearance of the aldehyde carbon signal and the presence of a new signal that corresponds to the carbon of an imine (Figure S4). The [3+6] stoichiometry of nanobarrel 1 was confirmed by MALDI‐TOF MS. The MS spectrum showed a molecular ion peak ([M+Ag]⁺) and isotopic distributions consistent with the molecular weight of nanobarrel 1 (Figure 2d). Most importantly, the peaks corresponding to other cages such as those with dimeric [2+4] or tetrameric [4+8] stoichiometries were not detected. This is consistent with DOSY analysis and also with the calculated reaction energies (after 0.1 ns molecular dynamics runs at the GFN2‐xTB level) for the [2+4] dimeric, [3+6] trimeric and [4+8] tetrameric nanobarrels that illustrate that the [3+6] trigonal‐prismatic barrel is the most thermodynamically stable (Figure S1 and Table S2).

Figure 2.

a) ¹H NMR spectra of precursor 2 and nanobarrel 1 and b) DOSY spectrum of nanobarrel 1 (y axis unit is cm² s⁻¹) and c) VT NMR spectra of nanobarrel 1 in 1,1,2,2‐tetrachloroethane‐d₂. Asterisks indicate solvent residual peaks. d) Experimental (MALDI‐TOF) and theoretical isotopic distributions of the [M+Ag]⁺ ion peak of nanobarrel 1.

All efforts to grow single crystals of nanobarrel 1 suitable for X‐ray diffraction were unsuccessful. To gain further insight into the structure, the conformational space of nanobarrel 1 was explored in a solvent continuum computed with eXtended Tight Binding (xTB) quantum chemistry Hamiltonian. From this, different low energy conformations were selected, optimized and energy ranked to be later refined with DFT. Final geometry optimizations were performed at the M06‐2X‐6‐31G(d,p) level in chloroform. The lowest energy molecular structure of nanobarrel 1 shows a symmetrical triangular prismatic structure in which three pyrene side walls are clipped by imine bonds through six (1R,2R)‐1,2‐diaminocyclohexane (3) units at the vertices. This arrangement generates two large triangular windows with a diameter of 12.7 Å (Figure 1b and c).

The UV/Vis spectrum of nanobarrel 1 (maxima at 295 and 356 nm) in chloroform is very similar to that of precursor 2 but slightly broadened and bathochromically shifted for 4 nm (Figure S5). The spectrum is dominated by the pyrene moieties and is vibronically resolved at the lowest energy transition. The absorbance ratio between the 0‐0 and 0‐1 vibronic bands (A_0‐0/A_0‐1) is 1.13 in both precursor 2 and nanobarrel 1, which indicates that there is no aggregation in solution. The emission spectrum of nanobarrel 1 in chloroform exhibits a significant hypochromic shift (40 nm) compared to that of precursor 2 (Figure S5), which was ascribed to the different functional groups in the phenyl substituents (imines or aldehydes).

Nanobarrel 1 is constituted by three pyrene panels and two large triangular windows with a diameter of 12.7 Å that are optimal to bind C₆₀ (Figure 3a–c). To investigate the ability of nanobarrel 1 to complex C₆₀, we performed simultaneously absorption and fluorescence titrations following a procedure that has been recently used to study the complexation between cages and fullerenes,[ ^8g , ^11b ] in which a solution of nanobarrel 1 (10⁻⁵ M) is titrated with a solution of C₆₀ (10⁻³ M). We used a 9 : 1 mixture of chloroform‐toluene in the titrations as it is able to dissolve both the host and the guest. In the absorption titrations, even if some changes were observed upon the addition of C₆₀, the band overlap in the UV region did not allow the estimation of the binding constant (Figure S6). In the fluorescence titrations, we could clearly observe how the fluorescence band of nanobarrel 1 at 416 nm is quenched, bathocromically shifted and broadened with the incremental addition of C₆₀ (Figure 3d). A binding constant (K _a) in the order of 10⁶ M⁻¹ was estimated for a 1 : 1 complex (the stoichiometry was confirmed by MS, see below) from three different titration experiments using either the Benesi–Hildebrand method ^[14] ((2.66±0.14)×10⁶ M⁻¹) (Figure 3e) or Bindfit's Nelder‐Mead method[ ¹⁵ , ¹⁶ ] ((3.39±0.48)×10⁶ M⁻¹).

Figure 3.

a) Molecular structure of C₆₀@1. b) Side and c) front views of the calculated structure of C₆₀@1. d) Fluorescence titration experiment of a solution of nanobarrel 1 (10⁻⁵ M) in chloroform‐toluene (9 : 1) with a solution of C₆₀ (10⁻³ M) in chloroform‐toluene (9 : 1). e) Benesi–Hildebrand fit for a 1 : 1 complex.

Given the high association constant and the high solubility of nanobarrel 1 in dichloromethane, chloroform and 1,1,2,2‐tetrachloroethane, in which C₆₀ is virtually insoluble, ^[17] we decided to investigate the ability of the nanobarrel 1 to solubilize C₆₀ in such solvents. When C₆₀ powder was added to a solution of nanobarrel 1 in dichloromethane, the color of the solution instantly changes from transparent to purple (Figure 4a). Simultaneously, the fluorescence of nanobarrel 1 is quenched upon the addition of C₆₀. This behavior was also observed in chloroform and 1,1,2,2‐tetrachloroethane and is consistent with the complexation of C₆₀ into the cavity of nanobarrel 1. To ensure a complete complexation, a 3‐equivalent excess of C₆₀ was added to a solution of nanobarrel 1 and the resulting solution was kept overnight at room temperature. Then the excess of C₆₀ was filtered and the solvent was evaporated. The formation of the complex C₆₀@1 was confirmed by NMR spectroscopy, MALDI‐TOF MS. The ¹H NMR spectrum shows that the aromatic signals of the C₆₀@1 complex are shifted in comparison to those of the free nanobarrel. These shifts are consistent with previous observations ^[18] and were ascribed to π‐π stacking between C₆₀ and the nanobarrel (Figure 4b). The DOSY spectrum of the C₆₀@1 complex shows a diffusion constant (9.78×10⁻⁷ cm² s⁻¹) that is virtually the same to that of nanobarrel 1 (Figures S7). This indicates that the size and geometry of the nanobarrel remains practically unaffected after complexation, hence that C₆₀ is complexed inside the cavity, in line with previous observations.[ ^8g , ^10u ] The ¹³C NMR spectrum of C₆₀@1 shows the characteristic signal at 141 ppm of C₆₀ that was not present in the spectrum of nanobarrel 1 (Figure 4c). The MS spectrum shows a molecular ion peak ([M+Ag]⁺) and isotopic distributions consistent with the molecular weight of C₆₀@1 (Figures 4d and S8). The UV/Vis electronic absorption spectra of C₆₀@1 and C₆₀ shows that the bands of C₆₀ in the C₆₀@1 complex (334 and 407 nm) are redshifted in comparison with those of free C₆₀ (329 and 405 nm), which is again consistent with π‐stacking between the nanobarrel and C₆₀ (Figure S9).

Figure 4.

a) Photographs of the solutions of nanobarrel 1, C₆₀ and C₆₀@1 in dichloromethane under exposure to natural and UV light. b) Partial ¹H NMR spectra of nanobarrel 1 and C₆₀@1 in (1,1,2,2‐tetrachloroethane‐d ₂) and c) ¹³C NMR (dichloromethane‐d ₂) spectra of nanobarrel 1 and C₆₀@1. Asterisk indicates solvent residual peaks. d) Experimental (MALDI‐TOF) and theoretical isotopic distributions of the [M+Ag]⁺ ion peak of C₆₀@1.

Similarly, we were not able to grow crystals of C₆₀@1 suitable for X‐ray diffraction. To gain additional insight into the structure of the C₆₀@1 complex, its geometry was computed at the M06‐2X‐6‐31G(d,p) level (Figure 3b, c). The model of C₆₀@1 illustrates the perfect fit between C₆₀ and the cavity of nanobarrel 1. For instance, the complexation of C₆₀ has virtually no influence in the geometry of the nanobarrel (Figures 3b, c), which is consistent with the DOSY spectrum of the C₆₀@1 complex that show almost the same diffusion rate as the empty nanobarrel 1 (Figures 2b and S7). Since the binding constant was experimentally estimated from a chloroform‐toluene (9 : 1) mixture, the binding energy was computed (M06‐2X‐6‐31G(d,p)) both in chloroform and in toluene, yielding similar values of −21.7 and −21.3 kcal mol⁻¹, respectively. These values compare favourably against the binding energy value of −8.8 kcal mol⁻¹ estimated experimentally. To further, investigate the effect of the solvent on the complexation the binding energy was calculated in vacuum with a value of −44.3 kcal mol⁻¹. The higher binding energy calculated in vacuum in comparison to that calculated in the presence of solvent illustrates the excellent fit between C₆₀ and the cavity of nanobarrel 1, but it also indicates that the nanobarrel is able to accommodate solvent molecules when empty, halving the binding energy in both chloroform and toluene and making it closer to the experimentally derived value.

The frontier molecular orbitals of C₆₀@1 computed at the M06‐2X‐6‐31G(d,p)‐chloroform level show three degenerate HOMOs and three degenerate LUMOs that are close in energy to the values for the HOMOs of nanobarrel 1 and the LUMOs of C₆₀, respectively (Figure 5a and Table S3). Consequently, in the C₆₀@1 complex, the electron density of the HOMOs is localized on the nanobarrel, whereas the electron density of the LUMOs is localized on C₆₀ (Figure 5b).

Figure 5.

a) Energy levels of C₆₀, nanobarrel 1 and C₆₀@1. b) HOMO (bottom) and LUMO (top) frontier molecular orbitals of C₆₀@1 at the M06‐2X‐6‐31G(d,p)‐chloroform level.

Conclusion

To conclude, we have described a new trigonal prismatic organic nanobarrel (1) that has been synthesized by dynamic covalent chemistry, using imine condensation between a carefully designed tetratopic pyrene precursor 2 and (1R,2R)‐1,2‐diaminocyclohexane (3). The structure of nanobarrel 1 has been confirmed by the ¹H NMR, ¹³C NMR, and MALDI‐TOF MS. The energy optimized molecular structure of nanobarrel 1 shows a symmetrical triangular prismatic structure with two large triangular windows with a diameter of 12.7 Å optimal to host C₆₀. Remarkably, nanobarrel 1 is able to encapsulate and solubilize C₆₀ in solvents in which C₆₀ is virtually insoluble, such as dichloromethane, chloroform and 1,1,2,2‐tetrachloroethane. The formation of the C₆₀@1 complex has been unequivocally confirmed by ¹H NMR, ¹³C NMR and MALDI‐TOF MS. Theoretical calculations show the perfect fit between nanobarrel 1 and C₆₀, in which the stabilization of C₆₀ within the cavity by π‐stacking has little influence in the geometry of the nanobarrel. The synthesis of this nanobarrel opens the door to the synthesis of new families of π‐extended nanobarrels and to the exploration of their optoelectronic, chiroptical, redox and host–guest properties in electronic and energy applications.

Conflict of interest

The authors declare no conflict of interest.

Supporting information

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supporting Information

Bera S., Das S., Melle-Franco M., Mateo-Alonso A., Angew. Chem. Int. Ed. 2023, 62, e202216540; Angew. Chem. 2023, 135, e202216540.

Data Availability Statement

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