Supramolecular Proton Conductors Self-Assembled by Organic Cages.

Proton conduction is vital for living systems to execute various physiological activities. The understanding of its mechanism is also essential for the development of state-of-the-art applications, including fuel-cell technology. We herein present a bottom-up strategy, that is, the self-assembly of Cage-1 and -2 with an identical chemical composition but distinct structural features to provide two different supramolecular conductors that are ideal for the mechanistic study. Cage-1 with a larger cavity size and more H-bonding anchors self-assembled into a crystalline phase with more proton hopping pathways formed by H-bonding networks, where the proton conduction proceeded via the Grotthuss mechanism. Small cavity-sized Cage-2 with less H-bonding anchors formed the crystalline phase with loose channels filled with discrete H-bonding clusters, therefore allowing for the translational diffusion of protons, that is, vehicle mechanism. As a result, the former exhibited a proton conductivity of 1.59 × 10-4 S/cm at 303 K under a relative humidity of 48%, approximately 200-fold higher compared to that of the latter. Ab initio molecular dynamics simulations revealed distinct H-bonding dynamics in Cage-1 and -2, which provided further insights into potential proton diffusion mechanisms. This work therefore provides valuable guidelines for the rational design and search of novel proton-conducting materials.

Introduction

Proton conduction plays an important role in the physiological activity of living systems.[1] The understanding of its mechanism is also essential for the development of state-of-the-art applications such as fuel-cell technology.[2,3] Over the last decade, high-crystallinity frameworks, including metal–organic frameworks (MOFs)[4−11] and covalent–organic frameworks (COFs),[12−18] have received considerable attention. These frameworks exhibit high proton conductivity up to 10–2 S cm–1, even paralleling the performance of Nafion.[19] Featuring tunable porosity and functionalizable channels, they provide various structural models for studying the effect of different parameters, such as porosity, functionality, and carrier concentration/mobility, on proton conduction. On the other hand, the large variety of these structural parameters also imposes formidable challenges for the clear elaboration of the correlation between these factors and the proton transport mechanism and conductivity.

Organic cages are a class of molecules with intrinsic nanosized cavity and rich structural diversity, which recently have attracted increasing attention.[20−31] By virtue of their excellent solution-processability, which are distinct from MOF and COF, they can self-assemble in solution into defect-free crystalline porous materials. Organic cages recently were applied for the preparation of solid-state proton conductors by the Cooper laboratory[32] and our group,[33] which hitherto had remained the only two examples for such a purpose. In this article, we would like to report on the rational design of two organic cages featuring different shapes, cavities, and number of proton carrier anchors. As a bottom-up strategy, their self-assembly leads to the formation of supramolecular proton conductors with different hydrogen-bonding architectures (H-bonding network vs cluster) and proton transport pathways, which can serve as powerful models for the in-depth study of these structural factors on proton conduction efficiency and mechanism (Grotthuss[34] vs vehicle[35] mechanism) at molecular and supramolecular levels, and provide guidelines for the rational design and search of novel proton-conducting materials (Figure ).

Figure 1

Schematic illustration of supramolecular proton conductors self-assembled by organic cages. By tuning the shape, cavity size, and number of proton carrier anchors, the H-bonding architecture (H-bonding network vs cluster, zoom-in view) and proton transport pathways of the proton conductors are precisely tailored, providing ideal models to elaborate structural correlation with the proton conduction mechanism and efficiency at molecular and supramolecular levels. H-bonds are shown in cyan dashed lines, and proton conduction pathways in yellow shading.

Results and Discussion

Rational Design and Synthesis of Self-Assembling Organic Cages

Aided by density functional theory calculations (Figure S1 in the Supporting Information), we first designed two organic cages, that is, Cage-1 and -2 with an identical chemical nature, but distinct shape, cavity size, and number of H-bonding anchors (Figure ). Amine groups on the cages were devised as hydrogen-bonding anchors to proton carriers, namely, trifluoroacetic acid (TFA) and water molecules in the current study (vide infra). The pendant hydroxyl group(s) on Cage-1 and -2 was introduced for solubility and easy solution processibility. As the molecular packing of supramolecular structures and their functions are determined by the self-assembling synthons,[36−42] we reasoned that the self-assembly of the two cages would yield two different supramolecular proton conductors with distinct intermolecular channels, which might alter the proton conduction behaviors of the encapsulated proton carriers.

Figure 2

(a) Synthetic scheme of Cage-1 and -2 by cycloimination between precursor 1 and TREN and TRPN molecules, respectively, followed by subsequent reduction. (b) Crystal structure of TFA-doped Cage-1 with large cavity size. Its eight nitrogen atoms are highlighted in blue. Six TFA and four H2O molecules are located inside and outside its cavity. (c) Crystal structure of TFA-doped Cage-2 with small cavity size. Its four nitrogen atoms are highlighted in green. Four TFA and two H2O molecules are all located outside its cavity.

Based on this rationale, Cage-1 and -2 were therefore synthesized by the cycloimination of aldehyde-containing precursor 1 and tris(2-aminoethyl)amine (TREN) and tris(3-aminopropyl)amine (TRPN), respectively, followed by the subsequent reduction of their imines into amine form (Figure a). Their purity and chemical structures were confirmed by high-performance liquid chromatography (HPLC), 1H and 13C nuclear magnetic resonance (NMR) spectroscopy, and matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS), and the detailed characterizations can be seen in the Supporting Information.

Crystallographic Analysis of TFA-Doped Cage-1 and -2

The single crystals of Cage-1 suitable for X-ray crystallography were obtained by slow evaporation of its hydrous tetrahydrofuran (THF) solution containing additional TFA, where H2O and TFA molecules were therefore introduced as proton carriers during the preparation of supramolecular proton conductors (Figure b). It crystallized into monoclinic space group C2/c and unambiguously verified the proposed structure of Cage-1. The cage adopts an irregular inner cavity with a relatively large size (ca. 1.8 nm × 1.1 nm), and two pendant hydroxyl groups are symmetrically distributed on its both sides. Six TFA and four H2O molecules are located inside and outside its cavity (Figures b and S8a). This composition was also confirmed by thermal gravimetric analysis (TGA) of the sample, as ca. 3 and 31% weight loss were observed, which correspond to 4H2O and 6TFA per Cage-1, respectively (Figure S13a).

As expected, it exhibits eight potential proton carrier anchors in the form of secondary and tertiary amines. However, the crystallographic analysis reveals that only six secondary amines on the aliphatic chains are protonated, which are bound with five CF3COO– anions and two H2O molecules by hydrogen bonding (Figure S8a), while one CF3COO– and two H2O molecules are unbound. The two tertiary amines are intact, which is presumably due to the electrostatic repulsion of the neighboring TFA anions and the steric hindrance of the irregularly stretched aliphatic moiety caused by the rigid strain conformation (Figure S8a).

The single crystals of Cage-2 suitable for X-ray crystallography were also obtained in the same way. It crystallized into the triclinic space group P1̅ (Figure c). As expected, it exhibits only four potential amino proton carrier anchors, and a much smaller inner cavity (ca. 0.6 nm × 1.0 nm) as compared to Cage-1. All four amines of the cage are protonated and are surrounded by four CF3COO– anions and two H2O molecules (Figures c and S8b). The composition was further confirmed by TGA of the sample, as ca. 3 and 32% weight loss were observed, which roughly correspond to 2H2O and 4TFA per Cage-2, respectively (Figure S13b). As the cavity of Cage-2 is too small to contain any host molecule, the proton carriers are all distributed outside its framework, where three CF3COO– anions and one H2O are anchored with the protonated amines by H-bonding, leaving one CF3COO– and one H2O unbound (Figure S8b).

Proton Hopping Pathways Formed by Dense H-Bonding Networks versus Proton Diffusion Channels Filled with Discrete H-Bonding Clusters

As the shape, cavity size, and the number of proton carriers per cage of Cage-1 and -2 were unambiguously clarified by single-crystal X-ray crystallography (SC-XRD), we then continued to examine the molecular packing of the cages within their TFA-doped crystalline phases, particularly probing the H-bonding architecture and proton conduction pathways within the intrinsic cavity and the intermolecular channels, if there is any, of the cages.

The self-assembly of TFA-doped Cage-1 molecules generates a closely packed crystalline phase. Figure a presents a zoom-in view with a cage-dimer. It shows four proton carriers, that is, two TFA and two water molecules reside within Cage-1, and a H-bonding network penetrates the intrinsic cavity of each cage, which is connected with the proton carriers located outside of the cage. As the packing of these molecules is so compact that no channels are available for apparent mass transfer, which might preclude the translational diffusion of proton carriers. When viewing the crystalline phase along crystallographic a-axis, four H-bonding networks (highlighted in yellow in Figure b) are formed within a diameter of D1 = 7.3 nm, which we coin proton hopping pathways. When cage frameworks are omitted but their nitrogen atoms are retained (Figure c), it clearly reveals this pathway, which is composed of a dense H-bonding network with a relatively narrow width (D2 = 1.1 nm).

Figure 3

Comparison of different H-bond (shown in dashed cyan lines) architectures and proton conduction pathways in the crystalline phase of TFA-doped Cage-1 and -2, respectively. The zoom-in view shows detailed information of (a) H-bond network vs (d) cluster, where cage frameworks are presented in the wireframe, TFA molecules in the ball–stick model, nitrogen atoms of the cage in blue for Cage-1, green for Cage-2, and H2O molecules in red balls, respectively. Highlighted in yellow are (b) four proton hopping pathways within the crystalline phase of TFA-doped Cage-1 with a diameter of D1 = 7.3 nm, and (e) three proton diffusion channels within the crystalline phase of TFA-doped Cage-2 with a diameter of D3 = 7.1 nm. (c) One pathway formed by dense H-bonding networks with a relatively narrow width (D2 = 1.1 nm) composed of nitrogen atoms (blue ball) of Cage-1 and oxygen atoms (red balls) of proton carriers (TFA and H2O), and (f) one channel filled with discrete H-bonding clusters within a relatively wide channel (D4 = 1.5 nm) formed by nitrogen atoms (green ball) of Cage-2 and oxygen atoms (red balls) of the same proton carriers. Solvent molecules (THF) and hydrogen atoms are omitted for clarity.

Contrary to the H-bonding networks in the crystalline phase of Cage-1, discrete H-bonding clusters are observed within the intermolecular channels formed by Cage-2 molecules, as shown by the zoom-in view with a cage-tetramer (Figure d). Besides, as highlighted in yellow, the self-assembly of Cage-2 molecules generates only three mass transfer channels within a diameter of D3 = 7.1 nm along the crystallographic a-axis (Figure e), which is less dense as compared to that of Cage-1. These H-bonding clusters are confined within the actual proton diffusion channel with a relatively larger width (D4 = 1.5 nm), suggesting a loose H-bond density as compared to the case of Cage-1 (Figure f).

Insights into the Proton Conductivity of the Supramolecular Conductors

As single crystals of sufficiently large size were unavailable, for probing the temperature-dependent conductivity, we therefore used the pelletized samples[11] of freshly prepared microcrystalline powder of TFA-doped Cage-1 and Cage-2, respectively. After we confirmed that the powder X-ray diffraction (PXRD) patterns of these microcrystalline samples are identical to those of the corresponding single crystals (Figures S9 and S10), we measured their conductivity by using alternating current electrochemical impedance spectroscopy (EIS) at different temperatures (303–353 K) under a fixing RH of 48% at which the crystals were formed. We first verified the necessity of the proton carrier TFA. Without doping with TFA, either Cage-1 or Cage-2 did not exhibit meaningful conductivity (Figure S14); while in the presence of TFA, the conductivity of their crystals was remarkably enhanced (Figure ).

Figure 4

Proton conductivity and electrochemical data for supramolecular proton conductors formed by Cage-1 and Cage-2. (a,b) Nyquist plots of TFA-doped crystals of Cage-1 and Cage-2 as a function of temperature (303–353 K) under air humidity [48% relative humidity (RH)]. (c) Arrhenius plots of the proton conductivity of Cage-1 (blue) and Cage-2 (green) under 48% RH. Least-squares fittings are shown as blue and green solid lines, respectively.

For the supramolecular proton conductor of TFA-doped Cage-1, the conductivity increases steadily as a function of temperature, from a relatively high value of 1.59 × 10–4 S/cm at 303 K to 3.75 × 10–4 S/cm at 353 K (Figure a and Table S3). In addition, its activation energy is 0.16 eV (Figure c), and this value lower than 0.4 eV indicates that the Grotthuss mechanism[34] is involved. This low activity energy is thought to be attributed to the confinement effect of the cage-like structure with the intrinsic cavity, which facilitates the fast proton migration aided by rapid intra-cage proton transfer.[32] It therefore suggests that the H-bonding networks are indispensable for proton hopping, where a proton donor releases and passes its proton to a neighboring accepter through the breakage and reformation of hydrogen bonding (Figure a,c).

On the other hand, the supramolecular proton conductor of TFA-doped Cage-2 shows a much lower conductivity of 9.19 × 10–7 S/cm at 303 K, which rises dramatically with the increase in temperature and reaches 1.04 × 10–5 S/cm at 353 K (Figure b and Table S3). Together with its calculated activation energy of 0.45 eV (>0.4 eV, Figure c), it implies the proton conduction proceeds via vehicle mechanism,[35] as in the absence of a continuous network of hydrogen bonds, the translational diffusion of proton carriers through the intermolecular channels formed by Cage-2 is required for the long-range proton conduction. This diffusion might also echo with the presence of disordered TFA only observed in the crystalline phase of Cage-2 by SC-XRD (Figure S8a vs S8b).

As revealed by PXRD after impedance measurements, the crystallinity of TFA-doped Cage-1 and Cage-2 were almost retained under a RH of 48% (Figures S11 and S12). By following the precedent protocols,[7,10] the cycle stability of the two materials were tested, showing that their proton conductivity remained stable for five cycles (Figure S19). When fixing the testing temperature at 303 K, we also examined the influence of RH (Table S3 and Figures S16 and S18). It reveals that the conductivity of TFA-doped Cage-1 steadily increases from 1.59 × 10–4 to 1.60 × 10–3 S/cm with the increase of RH from 48 to 98%, whereas it rises from 9.19 × 10–7 to 5.23 × 10–5 S/cm for TFA-doped Cage-2. It is also worth noting that the deterioration of the crystallinity of the two cages is somewhat observed after the impedance measurements with RH higher than 48%.

According to the Arrhenius equation, the proton conductivity is determined by both proton conduction mechanism and concentration of the proton carriers.[11] In the current study, the H-bonding architecture, that is, H-bond network versus cluster, is distinct for the crystalline phases of Cage-1 and Cage-2. It correlates well with the considerable dissimilarity of their activation energy, which in turn is related to the difference between Grotthuss and vehicle mechanism. Second, as the RH of 48% for the EIS experiments was kept identical to that during the crystal growth for Cage-1 and Cage-2, it is reasonable to assume that the density of proton carriers and transport pathways determined by SC-XRD and TGA can be used for direct comparison. As compared to Cage-2, the crystalline phase of Cage-1 exhibits a higher number of proton carriers per cage and a higher density of H-bonding pathways. In concert with the lower activation energy, these factors therefore lead to the higher performance of proton conduction of Cage-1.

To understand the atomistic mechanisms of proton diffusion in the crystalline phases of TFA-doped Cage-1 and Cage-2, we performed ab initio molecular dynamics (AIMD) simulations at a temperature of 500 K for up to 50 ps (see the ab initio molecular dynamics simulations section in the Supporting Information for computational details; the code used for data analysis is available from the GitHub repository: Molecular-Dynamics-Molecular-Cages). Due to the large sizes of the organic cage molecules and relatively low proton hopping rates (especially for Cage-2), we focused on the diffusion of water molecules. We first calculated the diffusion coefficients of water molecules (Dwater) in Cage-1 and Cage-2, and it reveals that water molecules diffuse much faster in Cage-1 (Dwater = 4.00 × 10–5 cm2 s–1) than Cage-2 (Dwater = 6.72 × 10–6 cm2 s–1, Figure a,b). The faster diffusion of water in Cage-1 than Cage-2 can be rationalized by H-bonding dynamics. We show the ratios between the number of hydrogen bond n(H-bonds) and the total number of unique hydrogen bonding donor–acceptor pairs N(d–a) as a function of AIMD simulation time in Figure c,d. For Cage-1, the mean n(H-bonds)/N(d–a) ratio is 0.0088, and for Cage-2, the mean n(H-bonds)/N(d–a) ratio is 0.0108, which is 23% higher than that of Cage-1. As water molecules act as both hydrogen donors and acceptors, the lower n(H-bonds)/N(D–A) ratio indicates the lower probability of the proton being trapped by water molecules due to H-bonding in Cage-1 than in Cage-2.

Figure 5

Comparison of average mean square-displacements (MSDs) and diffusion coefficients (Dwater) of water molecules for (a) Cage-1 and (b) Cage-2, the upper and lower bounds of MSDs and Dwater denoting their standard errors. The evolution of hydrogen bond numbers per unique available hydrogen bond donor–acceptor pairs in (c) Cage-1 and (d) Cage-2. The “window mean 50” in (c,d) denotes the corresponding smoothed data with a rolling window of width 50 × 0.005 ps.

We also calculated the H-bond lifetimes in Cage-1 and Cage-2 by using the time autocorrelation function[43,44] (see the Supporting Information for details), which shows that the average H-bond lifetimes in Cage-1 and Cage-2 are 2.99 and 5.36 ps (Figure S20), respectively. The shorter H-bond lifetime indicates that H-bonds in Cage-1 are on average less stable compared to those in Cage-2. This is in line with the frequent breakage and reformation of hydrogen bonding (Grotthuss mechanism) in the crystalline phase of TFA-doped Cage-1 and again implies that the proton may hop/diffuse faster in Cage-1 than Cage-2. Our AIMD simulations agree well with experimental observation of higher proton conductivity in Cage-1 than Cage-2. Based on the analysis of H-bonding dynamics in Cage-1 and Cage-2, we therefore highlight the importance of the rational design of the H-bonding architecture in organic cages to maximize the proton conductivity.

Conclusions

In summary, by taking advantage of the structural tunability and excellent solution processability of organic cages, we have developed a bottom-up strategy for the preparation of two supramolecular proton conductors with the characteristics of a superioninc conductor.[11,45] The two distinct supramolecular proton conductors were self-assembled by Cage-1 and -2 with an identical chemical nature, but different shape, cavity size, and H-bonding anchors. Their high-quality crystalline phases provide ideal models for the elaboration and direct comparison of proton conduction mechanisms. Large cavity-sized Cage-1 with more H-bonding anchors self-assembled into a supramolecular conductor with dense proton hopping pathways formed by H-bonding networks, where the proton conduction proceeded via the Grotthuss mechanism (Ea = 0.16 eV) and much higher proton conductivity up to 1.60 × 10–3 S/cm. Small cavity-sized Cage-2 with less H-bonding anchors formed a conductor with loose proton diffusion channels filled with discrete H-bonding clusters, therefore allowing for proton conduction via vehicle mechanism (Ea = 0.45 eV) but much lower conductivity (σ = 10–7 to 10–5 S/cm).

Due to the limited crystal size prepared from the two TFA-doped cages, this study employed the pelletized sample of microcrystals, whose grain boundary and size effect are yet to be clarified.[46] As a part of our continuous endeavor, we are currently trying to tether Brønsted acid/base pairs to the cage skeleton. By enhancing the electrostatic interaction, the growth of suitable-sized crystal from the cages can be envisioned, which might facilitate the investigation of the just-mentioned enquiries and the anisotropic proton conductivity of the crystals. Nonetheless, this work provides valuable guidelines for the rational design of novel proton-conducting materials. We also would like to show that cages can serve as promising building blocks for the search of novel supramolecular materials with emergent properties/functions, including but not limited to superconductors and ferroelectrics.[47]

Experimental Methods

Materials

Scandium(III) trifluoromethanesulfonate and sodium tris(acetoxy)hydroborate were purchased from Shanghai Adamas-Beta Co. Ltd. TREN, TRPN, 2-formylbenzeneboronic acid, and 1,4-phenylenediboronic acid were purchased from Beijing InnoChem Co. Ltd. All other reagents were bought from commercial sources and used without any purification unless otherwise stated. THF was dried over sodium/benzophenone under a nitrogen atmosphere before use. The reaction evolution was monitored by thin-layer chromatography, and flash column chromatography was performed on silica gel (200–300 mesh) with the indicated eluent.

General Instrumentation

NMR spectra were recorded on a Bruker Advance III HD (400/500 MHz) NMR spectrometer at room temperature. Fourier-transform infrared (FT-IR) spectroscopy spectra were recorded on a PerkinElmer Paragon 1000 spectrometer at frequencies ranging from 4000 to 500 cm–1 at room temperature. HPLC analysis was performed on a Shimadzu LC-20 AD instrument at room temperature using a Daicel Chiralcel IA column. MALDI-TOF MS was performed on a solariX XR 7.0 T hybrid quadrupole-FTICR mass spectrometer equipped with an ESI/APCI/MALDI ion source (Bruker Daltonics, Bremen, Germany). Single-crystal data were collected on a “Bruker APEX-II CCD” diffractometer (Ga Kα radiation, λ = 1.34139 Å, photon II detector). PXRD patterns were recorded on a Bruker D8 Advance diffractometer with Cu Kα1 radiation (λ = 1.5406 Å). TGA was performed with a STA449C integration thermal analyzer under flowing N2 with 10 °C/min ramp rate. The proton conductivity was measured with an EC Labs Bio-Logic (SP-300) potentiostat using banana plug cables.

Synthesis of Cage-1 and -2

Cage-1: into a 500 mL flask was charged with 0.005 mol/L triformyl precursor 1 (492.7 mg, 0.76 mmol, 1.0 equiv.) in 152 mL of CHCl3, and then, a 0.005 mol/L solution of TREN (214.5 mg, 1.2 mmol, 1.5 equiv) in 228.0 mL of CHCl3 was added dropwise, followed by the direct addition of Sc(OTf)3 (113.0 mg, 0.3 mmol, 0.3 equiv). The reaction mixture was stirred at room temperature for 4 h. Then, the product was reduced by NaBH(OAc)3 (2.4 g, 11.4 mmol, 15.0 equiv) overnight, and the solution was quenched with NaOH solution (2.0 M, 200 mL). The precipitates were filtered under vacuum and washed with dichloromethane (DCM) (100 mL), dried in the low-pressure oven to yield Cage-1 (315 mg, 56%) as a white solid. NMR spectroscopy was conducted with additional TFA for better solubility in CD3OD. 1H NMR (400 MHz, CD3OD): δ 7.63–7.20 (m, 52H), 6.75–6.73 (d, J = 8.0 Hz, 4H), 4.30 (s, 4H), 8.01 (s, 8H). 13C NMR (101 MHz, CD3OD): δ 156.87, 148.40, 148.23, 143.93, 143.88, 138.72, 138.65, 138.20, 132.95, 132.19, 132.09, 131.98, 131.02, 130.71, 129.99, 129.60, 129.44, 115.80, 65.38, 45.84. FT-IR (KBr, cm–1) ν: 3315.38 (br), 3023.87 (w), 2935.13 (w), 2827.78 (w), 1509.30 (s), 1578.44 (m), 1481.00 (m), 1408.00 (m), 1262.12 (m), 1178.54 (m), 1107.42 (w), 1006.19 (m), 830.23 (s), 763.59 (s), 707.11 (w), 649.53 (s), 597.90 (m), 539.48 (w). HR-MS (MALDI): C104H101N8O2+ [M + H]+ calcd, 1494.8076; found, 1494.8153. Elemental Anal. (%) Calcd for C104H100N8O2: C, 83.61; H, 6.75; N, 7.50. Found: C, 81.38; H, 7.07; N 6.87.

Cage-2: into a 500 mL flask was charged with 0.005 mol/L triformyl precursor 1 (492.7 mg, 0.76 mmol, 1.0 equiv) in 152 mL of CHCl3, and then, a 0.005 mol/L solution of TRPN (214.5 mg, 1.2 mmol, 1.5 equiv) in 228.0 mL of CHCl3 was added dropwise, followed by the direct addition of Sc(OTf)3 (113.0 mg, 0.3 mmol, 0.3 equiv). The reaction mixture was stirred at room temperature for 4 h. Then, the product was reduced by NaBH(OAc)3 (1.2 g, 5.7 mmol, 7.5 equiv) overnight, the excess reductant NaBH(OAc)3 was filtered off under vacuum, and the solution was quenched with NaOH solution (2.0 M, 200 mL), extracted with CHCl3 (3 × 200 mL), dried over anhydrous Na2SO4, and concentrated to give the crude product. Purification by flash column chromatography (DCM/MeOH/NH3(aq) = 50:2:3, v/v/v) afforded Cage-2 (312.0 mg, 52%) as a white solid. 1H NMR (400 MHz, CD2Cl2): δ 7.38 (d, J = 8.0 Hz, 6H), 7.26–7.32 (m, 18H), 7.20 (d, J = 8.0 Hz, 2H), 6.2 (d, J = 8.0 Hz, 2H), 7.36 (s, 6H), 2.43–2.47 (t, J = 16.0, 8.0 Hz, 6H), 2.26–2.30 (d, J = 16.0, 8.0 Hz, 6H), 1.27–1.41 (m, 6H). 13C NMR (101 MHz, CD2Cl2): δ 155.27, 146.79, 142.18, 139.26, 138.12, 132.84, 131.43, 130.79, 130.04, 128.53, 127.89, 127.70, 114.99, 64.08, 52.74, 52.26, 47.71, 27.04. FT-IR (KBr, cm–1)ν: 3024.2 (w), 2925.2 (s), 2852.8 (m), 1664.9 (w), 1609.2 (m), 1509.3 (s), 1481.0 (s), 1446.7 (m), 1376.7 (w), 1263.9 (s), 1178.3 (s), 1108.7 (m), 1006.1 (m), 828.4 (s), 762.1 (s), 737.3 (s), 597.0 (m), 570.5 (w), 537.6 (m). HR-MS (MALDI): C55H57N4O+ [M + H]+ calculated, 789.4527; found, 789.4511. Elemental Anal. (%) Calcd for C55H56N4O: C, 83.72; H, 7.15; N, 7.10. Found: C, 82.97; H, 7.24; N, 6.59.