Decoding Heptazine Architectures: From Molecular Association to Structural Insight

ABSTRACT

Heptazine (Hpz) derivatives display remarkable supramolecular organization governed by noncovalent interactions. Here, we explore how molecular chirality and structural variation influence the self‐association and organization of three Hpz derivatives (two chiral and one nonchiral) in solution and in the condensed phase. Variable‐temperature and concentration‐dependent NMR studies allowed the determination of dimerization and supramolecular aggregation constants, as well as the calculation of the thermodynamic parameters of these processes. Computational studies —including conformational analysis and molecular dynamics—corroborated the experimental findings, confirming the energetic stability and geometry of the proposed dimers. Moreover, photophysical studies showed aggregation‐induced emission (AIE) behavior, with blue light emission upon molecular association. Additionally, the introduction of chiral chains resulted in mirror‐image circular dichroism (CD) responses for the two Hpz enantiomers, demonstrating retention of chirality in bulk. In contrast, the nonchiral analogue displayed liquid crystalline (LC) behavior at high temperatures, characterized by a smectic mesophase. These results reveal how subtle structural variations modulate the supramolecular, optical, and mesomorphic properties of Hpz derivatives, highlighting the versatility of the heptazine core for multifunctional supramolecular and optoelectronic materials.

A comprehensive study (NMR and theoretical) of three novel heptazine derivatives that self‐assemble into dimers has been conducted, highlighting their versatility as a platform for multifunctional supramolecular and optoelectronic materials.

1. Introduction

Aromatic compounds with high nitrogen content, such as s‐heptazine or s‐tetrazine derivatives, have attracted considerable interest due to their exceptional structural and electronic characteristics [1, 2]. Specifically, the heptazine core presents a rigid and planar polycyclic aromatic framework with a particularly high nitrogen content, resulting in a low carbon‐to‐nitrogen ratio (C/N = 0.85). The presence of multiple nitrogen atoms strongly influences the distribution of the frontier molecular orbitals, stabilizing the LUMO and promoting spatial separation between the HOMO and LUMO orbitals [3]. As a consequence, s‐heptazine and its derivatives can exhibit distinctive optical and electronic properties, including strong electron affinity, high thermal and chemical stability, tunable fluorescence, and characteristic blue or blue‐green emission [1, 4] Furthermore functionalization of the heptazine core with π‐conjugated substituents or donor groups enables modulation of the HOMO‐LUMO gap and the emissive properties through increased electronic delocalization [3, 5, 6, 7]. These distinctive properties make heptazine‐based materials highly appealing for diverse applications such as optoelectronics [5], sensors [8, 9], photocatalysis [10, 11], gas storage [12, 13], and others [14, 15, 16].

Due to its flat geometry, rigid aromatic structure, and C3 symmetry, the s‐heptazine core has been traditionally considered as an excellent building block for columnar arrangements in discotic liquid crystals (DLCs) [17, 18, 19, 20, 21, 22]. In these systems, π–π stacking interactions between the aromatic cores promote the formation of columnar assemblies that lead to charge transport and anisotropic optical properties. Nonetheless, alternative mesophases can be also obtained through supramolecular association processes. Bala et al. described the first example of a smectic arrangement based on a dimer, in which the mesophase was induced by the H‐bonded ribbon‐like aggregation of the heptazine [23].

On the other hand, heptazine‐based materials have shown highly interesting luminescent properties in the aggregated or condensed state [5, 14, 24, 25]. Several functionalized s‐heptazine derivatives behave as strong blue emitters, whereas their fluorescence is considerably weaker in dilute solution. This behavior has been associated with aggregation‐induced emission (AIE) processes, in which aggregation restricts intramolecular motions and suppresses nonradiative deactivation pathways, thereby enhancing fluorescence in the solid or aggregated state [26, 27, 28]. Furthermore, the electron‐deficient nature of heptazine, in combination with electron‐donating substituents, provides an effective strategy for designing n‐type organic semiconductors [29, 30]. As a result, s‐heptazine derivatives have been successfully employed as AIE luminogens for solution‐processed OLEDs, exhibiting efficient deep‐blue emission in host‐guest configurations [18]. Notably, a few examples of liquid crystals that behave as AIE luminogens have been reported [31, 32, 33, 34, 35, 36, 37, 38, 39].

Herein, we describe three new heptazine‐derived molecules (Scheme 1) with structural and optical interesting properties. (R)‐ and (S)‐Hpz*‐C10 bear chiral centers, while the third Hpz‐C8 molecule is a nonchiral analogue. Surprisingly, these molecules exhibit a high degree of asymmetry in solution, as demonstrated through an in‐depth NMR study supported by theoretical calculations, and in which H bonds and π–π interactions stabilize the formation of dimers. For the compound Hpz‐C8, this asymmetry results in the formation of a smectic mesophase observed at high temperature.

SCHEME 1.

Chemical structure and proton and carbon assignments of heptazine derivatives (R)‐, (S)‐Hpz*‐C10 and Hpz‐C8.

2. Results and Discussion

2.1. Materials, Techniques, Preparation, and Characterization

Materials and methods used in this paper are gathered in Section 1 of the Supporting Information.

The synthetic route for the three covalent heptazine derivatives is illustrated in Scheme S1 in Section 2 of the Supporting Information and involves three main steps: preparation of cyameluric chloride, synthesis of aniline‐derived chains, and covalent coupling of these two components.

The synthetic procedures and characterization data are gathered in Section 2 and Figures S1–S6 of the Supporting Information. These molecules were characterized by High‐Resolution Mass Spectrometry (HRMS), Fourier Transform Infrared (FTIR), and Nuclear Magnetic Resonance (NMR) spectroscopies.

The elucidation of the structure of the heptazines described up to 1992 was mainly characterized by elemental analysis, melting points, and IR spectroscopic measurements [40]. In 1982, the first determination was made by X‐ray diffraction [41], and it was from 2002 onwards that the structures of simple heptazines began to be systematically described, providing in some cases ¹H NMR, ¹³C NMR, and ¹⁵N NMR data [1, 42]. Most NMR data were sharp spectra with well‐defined signals. Nevertheless, in the NMR spectra of heptazine derived from di‐ or tri‐arylamines, a splitting of the signals began to be observed both in ¹H and ¹³C NMR. Sayed et al. justified the aromatic ¹H splitting by the coupling of this proton with the NH of the amine, but without remarking on the appearance of new signals corresponding to the aromatic carbons [17]. Also in Bala's and Irla's works, complex spectra appeared in the data provided; however, they were not discussed either [19, 20]. Motivated by these unexplained features, we undertook a detailed NMR study aimed at understanding the origin of the new signals arising from the splitting of proton and carbon signals, which was subsequently completed with a theoretical study, including conformational analysis and molecular dynamics.

The assignment of signals in the ¹³C NMR spectra was carried out by 2D correlation experiments: ¹H‐¹³C HSQC (Heteronuclear Single Quantum Coherence) and ¹H‐¹³C HMBC (Heteronuclear Multiple Bond Correlation) (Figures S7 and S8). It is worth noting that C₁ and C₂ could not be assigned accurately because none of them showed any type of correlation with any other carbon or proton. Nonetheless, Jürgens et al. were able to assign them in melem in solid‐state NMR [42] and therefore, in this case, they should have similar positions. Regarding C₄, C₅, C₆, and C₇, due to their different chemical environments, they did not produce a single defined peak but rather a set of peaks within the same region. Surprisingly, in concentrated solutions of Hpz‐C8 in CDCl₃, shielding of the carbons of the Hpz core and, to a lesser extent, of the benzylic carbons C₅ and C₆ is observed, indicating a stacking of this heptazine in CDCl₃ solution (see Figure S9a).

Concentration‐dependent ¹H NMR studies and temperature‐dependent measurements were also performed for Hpz‐C8 and Hpz*‐C10 derivatives. These molecules form hydrogen‐bonded dimers in solution as deduced from dilution experiments: the upfield shift for the N‐H protons provides evidence for the involvement of these protons in H bonding. In addition, π‐π stacking of signals was also observed with increasing concentration. The ¹H NMR spectra recorded for Hpz*‐C10 at different concentrations in CDCl₃ are shown in Figure 1a, and a similar behavior was observed for the nonchiral heptazine (Figure S9b). Concentration‐dependent ¹H‐NMR spectroscopic data for all compounds can be found in the Supporting Information (Tables S1 and S2). At high concentrations, π‐π interactions promoted molecular stacking, leading to broader and less defined signals. As the concentration decreased, the signals corresponding to the H₅, H₆, and H₈ protons experienced deshielding due to the weakening of the Ring Current Effect (this behavior can also be observed in the protons belonging to the alkyl chains, see Figure S9c). The opposite outcome was observed for the NH protons (H₃), which shielded upon dilution, weakening the H‐bonding. Below 4 mM, the peaks resulted more defined, allowing for an improved signal assignment. The analysis of the downfield shift of the H₃ proton as a function of the Hpz concentration yielded small dimerization constants of K_dim = 20.7 ± 1.6 M⁻¹ for Hpz‐C8 and K_dim = 19.9 ± 0.7 M⁻¹ for Hpz*‐C10 [43]. Similarly, the aggregation constants have been calculated from the chemical shift data of H₆ considering an isodesmic model of supramolecular polymerization. This values were determined to be K_agg = 30.2 M⁻¹ ± 3.0 M⁻¹ for Hpz‐C8 and K_agg = 18.9 ± 2.1 M⁻¹ for Hpz*‐C10. (See Figure S10).

FIGURE 1.

Partial ¹H‐NMR spectra of Hpz*‐C10 (CDCl₃, 500 MHz) at different (a) concentrations at 298K and (b) temperatures at a concentration of 20 mM.

Temperature‐dependent measurements of ¹H signals were also performed. As observed in Figures 1b and S11, increasing the temperature led to remarkably sharper peaks. This improvement was attributed to an enhanced molecular motion at higher temperatures, which reduces signal broadening. Additionally, high temperatures weakened the molecular aggregation or π‐π stacking, leading to shifts of aromatic and alkoxyl protons toward higher frequencies, while the weakening of H‐bonds shifted the NH proton toward lower frequencies (See Table S3 in the Supporting Information). The variation of the equilibrium constants with temperature is represented by the van't Hoff equation, which allows the thermodynamic parameters of the dimerization process to be determined by a linear regression fit (Figure S12). Thus, the standard enthalpy and entropy were determined to be ΔH = ‐22.2 ± 1.0 kJ mol⁻¹ and ΔS = ‐49.9 ± 3.3 J mol⁻¹ K⁻¹ for Hpz‐C8 and ΔH = ‐29.7 ± 0.5 kJ mol⁻¹ and ΔS = ‐74.6 ± 1.5 J mol⁻¹ K⁻¹ for (R)‐ /(S)‐Hpz*‐C10. And for the supramolecular polimerizarion, the values were ΔH = ‐11.4 ± 0.5 kJ mol⁻¹ and ΔS = ‐8.8 ± 1.6 J mol⁻¹ K⁻¹ for Hpz‐C8 and ΔH = ‐12.8 ± 0.5 kJ mol⁻¹ and ΔS = ‐17.6 ± 1.6 J mol⁻¹ K⁻¹ for (R)‐/ (S)‐Hpz*‐C10. The negative enthalpy and entropy values indicate that these processes are enthalpy‐driven [44].

All these data seem to confirm that, even though the s‐heptazine core is flat and has a C3 symmetry, dimer formation hinders free rotation in solution, and the 2,5,8‐tris(4‐R‐phenylamino)‐s‐heptazine derivatives adopt a nonsymmetric conformation in which the nuclei are in different environments and are not magnetically equivalent, presenting complicated ¹H and ¹³C NMR spectra.

2.2. Theoretical Studies

To better understand the spatial arrangement of the Hpz‐C8, theoretical studies were conducted. A conformational search was performed (see Computational Details in Section 4 of the Supporting Information), leading to a set of 1000 conformers for Hpz‐C8. The most stable structures (with a relative energy with respect to the global minimum lower than 1.0 kcal/mol as calculated with the semiempirical PM‐7 method) shared certain features, such as a fully planar C‐N core, a single flexible independent arm with a high degree of freedom, or paired arms displaying strong linearity, as shown in Figure S13a. Notably, a very flat and aligned front between the C‐N core and one of the arms was observed. This structural feature will become particularly relevant in later discussions. Contrarily, significantly less stable conformations (with a relative energy higher than 6.5 kcal/mol at the PM‐7 level, see Figure S13b), while also showing a flat C‐N core, exhibited a significantly more pronounced globular geometry, and, in some cases, perpendicular aniline rings with respect to the C‐N core.

Regarding the possibility of dimer formation (Figure 2a), explicit molecular dynamics simulations of the Hpz‐C8 in a periodic box with a concentration of 55 mM in CHCl₃ revealed stability over time of Hpz‐C8 dimers (for a 10 ns simulation). Additionally, the hydrogen bond distances between the two Hpz‐C8 molecules were monitored during the simulation, revealing the presence of bifurcated interactions involving the N‐H protons. As shown in Figure 2b, the distance of the primary H‐bond remained relatively stable, with an average value of approximately 2.0 Å, indicating the presence of strong and well‐defined H‐bonding interactions. In contrast, the secondary interaction with a nearby nitrogen atom showed a longer average distance of around 2.7 Å. This difference in bond length is typical of bifurcated H‐bonds, where the primary interaction is stronger and more linear than the secondary one. The distance between the central nitrogen atoms of the dimerized Hpz‐C8 also remained relatively stable over time, with a value of approximately 7.5 Å (Figure 2b); and the length between C8 atoms (in the O‐CH₂ of each branch) being about 8.0 Å (Figure 2c). This result suggests that dimers are stable over time.

FIGURE 2.

(a) Dimer structure. (b) Evolution of the intermolecular N···H hydrogen‐bond distances during the 10 ns molecular dynamics simulation. (c) Evolution of the intermolecular C···C distance (in orange and cyan) between the O‐CH₂‐alkyl carbon atoms of the dimer during the 10 ns molecular dynamics simulation. (d) Representative configuration of stacked dimers obtained from the molecular dynamics simulation, showing the intermolecular stacking distance.

An additional molecular dynamics simulation was carried out for the stacking of Hpz‐C8 dimers. Throughout the 10 ns of simulation, the hydrogen bonds persisted relatively stable, while stacking remained in a diagonal orientation (as shown in Figure 3d), not perfectly vertical, at distances between approximately 3 and 6 Å (Figure S14a). Furthermore, the length of the intermolecular hydrogen bonds between adjacent Hpz‐C8 molecules forming a dimer showed a relatively stable value of about 2.0‐2.5 Å (Figure S14b), similar to the values obtained for the single dimer.

FIGURE 3.

POM images of Hpz‐C8 showing a SmA mesophase in the first cooling process at 240°C. (a) x10 and (b) x20.

2.3. Phase Behavior

The thermal properties of the compounds were studied by thermogravimetry (TGA), which showed that the temperature at which a 2% mass loss is reached was clearly higher than their isotropic points (Table 1). Thermal transitions and liquid crystalline (LC) properties were studied by differential scanning calorimetry (DSC) and polarizing optical microscopy (POM).

TABLE 1.

Thermal and thermodynamic properties of Hpz derivatives.

Compound	T2% (°C) a	Phase transition (T °C, [ΔH kJ/mol]) b
Hpz‐C8	317	I 262 [3.04] SmA 232 [26.3] Cr Cr 266 [38.1] I
(R)‐Hpz*‐C10	313	I 209 [25.7] Cr Cr 233 [28.1] I
(S)‐Hpz*‐C10	310	I 205 [26.8] Cr Cr 229 [29.9] I

^a Temperature at which 2% mass loss is detected in the thermogravimetric curve.

^b DSC data of the first cooling and second heating scans at a rate of 10°C min⁻¹. Temperatures (°C) are read at the maximum of the corresponding peaks, and enthalpies (kJ mol⁻¹) are given in brackets. Cr: cristal, I: isotropic liquid, SmA: smetic A mesophase.

After the first heating cycle, DSC curves also showed a good reproducibility in the following DSC heating‐cooling cycles (Figure S15). Both (R)‐ and (S)‐Hpz*‐C10 exhibited similar behavior but did not display mesogenic properties, unlike what was observed in the compounds described by Pal et al. [22], probably because the asymmetric C atoms are further away from the heptazine core. In contrast, Hpz‐C8 exhibited a likely disordered smectic arrangement at high temperatures (between 230 and 270°C). This notable difference is likely due to the pronounced aggregation effects observed in Hpz‐C8. It is important to emphasize that this higher aggregation tendency in Hpz‐C8 arises from its lack of asymmetric centers, allowing the molecules to stack more easily without the disruptions typically caused by chiral centers. Under POM, Hpz‐C8 exhibited a colorful fan‐shaped texture at high temperatures (>230°C) (Figure 3). Upon cooling, a new, more ordered smectic phase appeared, crystallizing later at lower temperatures, as also observed in the XRS measurements (vide infra).

2.4. Mesophase Determination by XRS

Since only Hpz‐C8 displayed mesogenic behavior, temperature‐dependent XRS measurements were only performed for this molecule. The virgin sample was first heated to the isotropic temperature and cooled down, next extruded at 90°C. The aligned fibre was then transferred to a capillary, sealed and glued, into the metal sample holder and it was measured at different temperatures.

Upon heating, Hpz‐C8 was first measured at 259°C, revealing the presence of a crystalline phase. Then, at 264°C, Hpz‐C8 had already transitioned to its isotropic point. Upon cooling to 250°C, the XRS measurement of Hpz‐C8 exhibited a pattern consistent with a smectic A phase (Figure S16 in the Supporting Information). The first order reflection of the lamellar stacking (0 0 1) indicated an average distance between layers of d ₀₀₁ = 27.2 Å. Using this interlayer spacing and the full width at half maximum of the (0 0 1) peak, the correlation length (ξ) was calculated to be 85.2 Å, which corresponds to approximately three correlated layers. This relatively small value suggests that, at this temperature, the lamellar structure reveals only a short‐range order. In the high‐angle region, the diffuse halo was found around 5.0 Å, consistent with the liquid‐like lateral packing of the mesogens in the layer. At 224°C, the XRS pattern showed overlapping peaks. Based on the POM observations, which indicated a transition toward a more ordered arrangement, the signal of SmA phase might be superimposed on that of the emerging new mesophase. To extract structural parameters, a deconvolution of the peak was performed, resulting in two well‐defined peaks (Figure 4). For the first peak (1), the interlayer spacing and its corresponding correlation length were d ₀₀₁ = 29.7 and ξ = 376 Å (12‐13 layers), respectively, while for the second (2), they were d ₀₀₁ = 27.3 Å and ξ = 57.6 Å (2‐3 layers). These results suggest that peak 1 might correspond to a higher‐order smectic A phase, whereas peak 2 matches a similar smectic A phase as the one observed at 250°C. At 190°C and below, the XRS measurements of Hpz‐C8 exclusively exhibited crystalline behavior, consistent with the POM observations. All structural data extracted from XRS measurements can be found in the Supporting Information (Table S4).

FIGURE 4.

Partial XRS pattern for Hpz‐C8 at 224°C after deconvolution, resulting in two well‐defined peaks 1 and 2.

2.5. Photo‐Physical Properties and AIE Behavior

In order to study the photophysical behavior of these heptazine compounds, UV‐vis and fluorescence spectroscopy were performed in solution (10⁻⁵ M in dichloromethane) and the solid‐state (Table S5 in the Supporting Information). Regarding absorption properties in solution, Hpz‐C8 displayed three peaks at 276, 302, and 325 nm. The chiral derivatives (R)‐Hpz*‐C10 and (S)‐Hpz*‐C10 showed three peaks at 278, 302, and 320 nm (Figure 5a). These peaks are likely attributable to π–π* and n–π* transitions. Their emission spectra, however, exhibited only one peak centered at 444 and 448 nm, respectively, for Hpz‐C8 and for (R)/(S)‐Hpz*‐C10. Stokes shift, the difference between the absorption and emission maxima, was calculated as 119 and 128 nm. In the solid state, all displayed two well‐resolved peaks at 286 and 329 nm for Hpz‐ C8, 286 and 326 nm for (R)‐Hpz*‐C10 and 287 and 330 nm for (S)‐Hpz*‐C10 (Figure 5b) The emission spectra revealed a maximum centered at 449 nm for Hpz‐C8, 439 nm for (R)‐Hpz*‐C10 and 437 nm for (S)‐Hpz*‐C10. All these values are in good agreement with those reported in the literature [17, 18, 19, 23].

FIGURE 5.

UV‐vis and emission spectra for (R)‐Hpz*‐C10 (a) in solution (DCM, 1×10⁻⁵ M) and (b) in solid. (c) (R)‐Hpz*‐C10 solutions display a blue color under 365 nm UV light. (d) Emission intensity of (R)‐Hpz*‐C10 with different water fractions from 0% to 90%.

Interestingly, heptazine derivatives usually exhibit AIE phenomena with blue‐green color emission [18, 22]. To study the AIE behavior, different fractions of water were added into the solution of Hpz‐C8 or (R)/(S)‐Hpz*‐C10 in THF (Figure 5c). As increasing the water fraction (f_w ), the aggregation of the compound intensifies. At 0% water fraction and up to 50%, almost no emission was found. Nonetheless, the intensity increased dramatically at f_w 70% water. At 90% f_w , the maximum emission was observed (Figures 5d and S17).

2.6. Ellipticity

As observed in NMR studies, this type of molecule tends to aggregate through π–π stacking interactions. However, despite not exhibiting mesogenic behavior, ellipticity measurements were recorded for each enantiomer. Samples were prepared on quartz substrates and rotated in the plane at each 60° to exclude the influence of linear dichroism and linear birefringence. As shown in Figure S18, both (R)‐Hpz*‐C10 and (S)‐Hpz*‐C10 displayed similar behavior, each showing a single band of opposite sign. The maxima appeared at 346 nm for the (R)‐Hpz*‐C10 and at 350 nm for (S)‐Hpz*‐C10. This small difference in wavelength maxima and intensity may be attributed to the purity of the initial enantiomers. Furthermore, as expected, the intensity of the bands was low, as these molecules do not form columnar structures and thus cannot generate a good transmission of chirality.

3. Conclusions

In summary, we demonstrated the formation of stable dimeric species within heptazine systems and elucidated their spatial organization through combined experimental and theoretical approaches. Variable‐temperature and concentration‐ dependent NMR studies enabled the determination of association constants and the corresponding thermodynamic parameters, revealing negative enthalpy and entropy values that indicate an enthalpy‐driven dimerization process stabilized by π–π stacking and H‐bonding interactions. Molecular dynamics simulations supported the proposed dimeric structures, confirming their energetic stability and providing molecular‐level insight into the association mechanism. Furthermore, photophysical investigations showed AIE properties, with blue light emission emerging upon molecular association. Remarkably, the nonchiral derivative Hpz‐C8 exhibited LC behavior at high temperatures, characterized by a smectic A mesophase, whereas the introduction of chiral centers led to mirror‐image CD responses, confirming the maintenance of chirality in the solid state for both Hpz*‐C10 enantiomers. Overall, these findings demonstrate how subtle structural variations govern the supramolecular, optical, and mesomorphic properties of heptazine derivatives, underscoring the versatility of the heptazine core as a platform for multifunctional supramolecular and optoelectronic materials.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

The authors have cited additional references within the Supporting Information [45, 46, 47, 48, 49, 50, 51, 52, 53].

Supporting File 1: chem71350‐sup‐0001‐SuppMat.docx.