Mechanochemical Syntheses of Isostructural Luminescent Cocrystals of 9-Anthracenecarboxylic Acid with two Dipyridines Coformers

Tuning and controlling the solid-state photophysical properties of organic luminophore are very important to develop next-generation organic luminescent materials. With the aim of discovering new functional luminescent materials, new cocrystals of 9-anthracene carboxylic acid (ACA) were prepared with two different dipyridine coformers: 1,2-bis(4-pyridyl)ethylene and 1,2-bis(4-pyridyl)ethane. The cocrystals were successfully obtained by both mechanochemical approaches and conventional solvent crystallization. The newly obtained crystalline solids were characterized thoroughly using a combination of single crystal X-ray diffraction, powder X-ray diffraction, Fourier-transform infrared spectroscopy, differential thermal analysis, and thermogravimetric analysis. Structural analysis revealed that the cocrystals are isostructural, exhibiting two-fold interpenetrated hydrogen bonded networks. While the O–H···N hydrogen bonds adopts a primary role in the stabilization of the cocrystal phases, the C–H···O hydrogen bonding interactions appear to play a significant role in guiding the three-dimensional assembly. Both π···π and C-H···π interactions assist in stabilizing the interpenetrated structure. The photoluminescence properties of both the starting materials and cocrystals were examined in their solid states. All the cocrystals display tunable photophysical properties as compared to pure ACA. Density functional theory simulations suggest that the modified optical properties result from charge transfers between the ACA and coformer molecules in each case. This study demonstrates the potential of crystal engineering to design solid-state luminescence switching materials through cocrystallization.


Introduction
Organic solid-state luminescent materials have been studied widely over the last decade [1][2][3]. This attention owes to their appealing and versatile optoelectronic applications in the fields of laser technology [4], chemical/biological sensors [5], and for organic light-emitting diodes [6]. The bulk fluorescent properties of solids are well-known to depend on the primary molecular structure of the individual components. Moreover, solid-state fluorescence properties are also influenced Scheme 1. Molecular structures of 9-anthracenecarboxylic acid (ACA) and different dipyridines coformers in the present study.
Following ball milling, the resulting powders were analyzed immediately by powder X-ray diffraction.

Preparation of Single Crystals of the Cocrystals
ACA (222 mg, 1 mmol) and BPEE (91 mg, 0.5 mmol) or BPE (92 mg, 0.5 mmol) in a 2:1 stoichiometric ratio were ground in a mortar with a few drops of MeOH for about 5 min. This procedure was repeated three times. The dry powder was subsequently dissolved in hot methanol (45 °C, 25 ml) and left to crystallize. Following approximately 4 days, block-shaped single crystals of both cocrystals were harvested and analyzed by single-crystal X-ray diffraction. Elemental analysis: ACA-BPEE cocrystal (1)
Following ball milling, the resulting powders were analyzed immediately by powder X-ray diffraction.

Preparation of Single Crystals of the Cocrystals
ACA (222 mg, 1 mmol) and BPEE (91 mg, 0.5 mmol) or BPE (92 mg, 0.5 mmol) in a 2:1 stoichiometric ratio were ground in a mortar with a few drops of MeOH for about 5 min. This procedure was repeated three times. The dry powder was subsequently dissolved in hot methanol (45 • C, 25 ml) and left to crystallize. Following approximately 4 days, block-shaped single crystals of both cocrystals were harvested and analyzed by single-crystal X-ray diffraction. SCXRD data for each cocrystal product were collected using a Bruker D8 Venture diffractometer (Bruker AXS, Karlsruhe, Germany) equipped with graphite-monochromated Mo Kα radiation (λ = 0.71073 Å). Data reduction was performed using the Bruker AXS SAINT [39] and SADABS [40] software packages. Each structure was solved through the use of SHELXS 2018 [41] using the direct method, followed by successive Fourier and difference Fourier synthesis. Full-matrix least-squares refinements were done on F2 using SHELXL 2018 [41], including anisotropic displacement parameters for all nonhydrogen atoms. During refinement of the ACA-BPE cocrystal, disorder of the carbon atoms (C12A and C12B) of BPE exhibited disorder; their occupancies were therefore fixed each at 0.50, before final refinement. Hydrogen atoms directly bonded to oxygen were located from the electron density maps, and all the hydrogen atoms bonded directly to carbon were fixed at their ideal positions. All additional calculations were performed within the SHELXS 2018, SHELXL 2018, and WinGX (Ver. 1.80) [42] software suites. Mercury v3.6 [43] was used to visualize and display structural figures. Data collection, structure refinement parameters, and crystallographic data of the cocrystals are summarized in Table 1. Prior to analysis, samples of BPEE and BPE were individually ground and packed into borosilicate capillaries (inner diameter 0.5 mm). PXRD data were collected from these capillary samples on a Bruker D8 Discover diffractometer (Bruker AXS, Karlsruhe, Germany) equipped with a LYNXEYE XE detector and Cu-K α1 radiation (λ = 1.506 Å). Data were collected over a range of 2θ = 5 • to 60 • with step size of 0.009 • . The time per step was 2.3 s. A ground sample of ACA was measured using a D8 Advanced diffractometer (Bruker AXS, Karlsruhe, Germany) in Bragg-Brentano geometry equipped with a LYNXEYE XE-T detector and a copper anode (Cu-K α , λ = 1.5406 Å). Data were collected over scattering angles 2θ = 5 • to 80 • with step size 0.02 • and 2 s per step. PXRD patterns were collected for all cocrystals between 4 and 40 • with a step size of 0.02 • and an accumulation time of 1 s per step.

Elemental Analyses
Carbon, hydrogen, and nitrogen analyses were performed using a Perkin-Elmer (Rodgau, Germany) 240C elemental analyzer.

Fourier-Transform Infrared Spectroscopy (FTIR)
FTIR spectra for powdered samples of each material were measured using a Nicolet FT-IR NEXUS (Thermo Fischer, Dreieich, Germnay) spectrometer equipped with a Diamond-ATR-Golden Gate unit and a DTGS KBr detector. Each spectrum was collected across a range of 200-4000 cm −1 with a spectral resolution of 4 cm −1 .

Raman
Raman spectra were recorded on a RXN1 TM analyzer (Endress+Hauser Process Analysis, Saint-Priest, France) with an excitation wavelength of λ = 785 nm. For the measurements, a con-contact probe head with the working distance of 10 mm and a spot size of 1 mm was used.
2.4.6. Thermal Analysis (Differential Thermal Analysis-Thermogravimetric (DTA-TG)) A simultaneous DTA-TG analysis was conducted using a TAG24 thermoscale (Setaram, Mougins, France). Measurements were performed in open platinum crucibles under continuous Ar-air flow. Measurements were conducted at a heating rate of 10 K/min 25 • to 500 • .

Photoluminescence Spectroscopy
Photoluminescence measurements were performed with an Edinburgh Instruments (Livingston, UK) FLS 980 fluorescence spectrometer. All photoluminescence spectra were measured in an integrating sphere. The samples were excited with a 450-W ozone-free xenon arc lamp. The detector was a R928P PMT, electrically cooled to −20 • C.

Computational Methods
Initial structures were taken from experimentally determined X-ray diffraction data, as described above. Plane wave density functional theory (pw-DFT) calculations were performed using Quantum ESPRESSO v6.4. [44,45]. Structure relaxation was performed using the exchange-correlation functional of Perdew-Burke-Ernzerhof (PBE) [46] with the exchange-dipole moment (XDM) [47] dispersion correction and damping parameters a 1 = 0.3275 and a 2 = 2.27673. Using the projector augmented-wave method, the electronic wave function was expanded in plane waves to a kinetic energy cut-off of 60 Ry for the wave function and 400 Ry for the charge density. Convergence of the electronic wave function was accepted <10 −8 Ry, and forces were considered as converged when <10 −3 Ry/a.u. The electronic band structures were subsequently calculated at the relaxed geometries using the screened hybrid functional HSE06 [48]. The electronic density of the states was plotted across no fewer than 200 k-points for each cocrystal.
Time-dependent density functional theory (TD-DFT) calculations were performed using Orca v4.2. [49]. All atoms were modeled at the B3LYP/def2-TZVP level. The resolution of identify approximation with def2/J auxiliary basis sets was used for the Coulomb integrals and COSX numerical integration for the Hartree-Fock exchange [50,51]. The asymmetric unit was extracted from the relaxed crystal structures, which was used to calculate the initial excitation energies. The interaction geometry between the dimers was fixed across the hydrogen bonding interactions, followed by relaxation of the dimer on the first singlet excited state potential energy surfaces. The vertical excitation energies were calculated in the relaxed state and taken as indicative of the photoluminescence output. Hydrogen bonding interaction energies were calculated within the framework of the symmetry adapted perturbation theory (SAPT), as implemented in PSI4 [52]. The "bronze-level" SAPT method was selected, as identified by Parker et al. [53]: sSAPT0/jun-cc-PVDZ. This method yields an error of ca 2.97 kJ/mol for hydrogen-bonded systems.

Result and Discussion
9-Anthracenecarboxylic acid (ACA) was screened for new multicomponent fluorescent solids with two different dipyridine coformers, 1,2-bis(4-pyridyl)ethylene (ACA-BPEE 1) and 1,2-bis(4-pyridyl)ethane (ACA-BPE 2). The screening was performed by mechanochemical reaction. Single crystals were grown by dissolving the ground powder in hot methanol. Both solids were obtained in quantitative yields, and their formation was substantiated by X-ray diffraction and spectroscopic and thermal methods of analyses. Crystallographic parameters are given in Table 1, and hydrogen bond lengths, along with the ORTEP diagrams of both multicomponent solids, are shown in the Supplementary Information. FTIR and Raman spectra of 1−2 are unique compared to the staring materials, supporting the successful formation of multicomponent solids. Moreover, the spectral similarities shared between the cocrystals suggest both solids share similar hydrogen-bonding patterns (Figures S1-S4 in the Supplementary Materials). In FTIR, the characteristic ν(C=O) band of ACA at 1668 cm −1 is blue-shifted to 1697 cm −1 in both the cocrystals (Figures S1 and S2). The Raman bands at 1408 and 1559 cm −1 for ACA, which correspond to C-C and/or ring deformation modes, are shifted in the cocrystals to 1412 and 1561 cm −1 , respectively ( Figure S3).

Mechanochemical Synthesis
To synthesize bulk ACA cocrystals, a mechanochemical strategy was employed by both neat grinding (NG), as well as liquid-assisted grinding (LAG), using methanol at 50 (Figure 2b for ACA-BPEE and Figure 3b for ACA-BPE). The cocrystals therefore appeared stable under the mechanical treatment. Furthermore, the addition of small amounts of methanol led to significant improvements in the reaction time.

Thermal Analysis
The thermal behavior of our newly synthesized multicomponent solids was investigated by differential thermal analysis (DTA) with coupled thermogravimetric (TG) analysis ( Figure 4). Literature melting points of the pure starting materials are ACA (213-217 °C), BPEE (148-152 °C), and BPE (110-112 °C). The corresponding experimental values obtained in this work are provided in Table S3. The DTA thermograms for both cocrystals exhibit a single well-defined endotherm. This is taken to be the melting point of each material. The DTA thermogram of the ACA-BPEE cocrystal exhibits a sharp endotherm with onset 217 °C corresponding to the melting, followed by decomposition of the material with a pronounced mass loss shown by the TG scan at ca. 280 °C (Figure 4a). The ACA−BPE cocrystal exhibits its melting endotherm at 190 °C , followed by decomposition at ca. 270 °C (Figure 4b). The melting point of ACA−BPE sits between those of the starting materials, whereas ACA−BPEE has virtually the same melting point as ACA (Table S3). In both cocrystals, a negligible weight loss was noticed prior to melting in all TG curves, which clearly indicates that all the multicomponent solids are un-solvated crystal forms. We do note the presence of a broad exotherm in both cocrystals, centered around ca. 65 °C. The origin of this conserved feature is presumably related to the ACA component in all the multicomponent solids ( Figure S7).

Thermal Analysis
The thermal behavior of our newly synthesized multicomponent solids was investigated by differential thermal analysis (DTA) with coupled thermogravimetric (TG) analysis (Figure 4). Literature melting points of the pure starting materials are ACA (213-217 • C), BPEE (148-152 • C), and BPE (110-112 • C). The corresponding experimental values obtained in this work are provided in Table S3. The DTA thermograms for both cocrystals exhibit a single well-defined endotherm. This is taken to be the melting point of each material. The DTA thermogram of the ACA-BPEE cocrystal exhibits a sharp endotherm with onset 217 • C corresponding to the melting, followed by decomposition of the material with a pronounced mass loss shown by the TG scan at ca. 280 • C (Figure 4a). The ACA−BPE cocrystal exhibits its melting endotherm at 190 • C, followed by decomposition at ca. 270 • C (Figure 4b).
The melting point of ACA−BPE sits between those of the starting materials, whereas ACA−BPEE has virtually the same melting point as ACA (Table S3). In both cocrystals, a negligible weight loss was noticed prior to melting in all TG curves, which clearly indicates that all the multicomponent solids are un-solvated crystal forms. We do note the presence of a broad exotherm in both cocrystals, centered around ca. 65 • C. The origin of this conserved feature is presumably related to the ACA component in all the multicomponent solids ( Figure S7).  (Figure 4b). The melting point of ACA−BPE sits between those of the starting materials, whereas ACA−BPEE has virtually the same melting point as ACA (Table S3). In both cocrystals, a negligible weight loss was noticed prior to melting in all TG curves, which clearly indicates that all the multicomponent solids are un-solvated crystal forms. We do note the presence of a broad exotherm in both cocrystals, centered around ca. 65 °C. The origin of this conserved feature is presumably related to the ACA component in all the multicomponent solids ( Figure S7).

Luminescent Property
Altering the crystallographic packing and intermolecular interactions of the organic luminophore can greatly influence the solid-state emission properties of a material. ACA itself is a well-known luminophore, and its luminescent properties were studied in the solution, as well as in the solid-state [54,55]. ACA therefore acts as an excellent candidate to study the effects of cocrystallization on luminescent properties.
To collect the photoluminescence spectra (PL), the PL excitation (PLE) spectra for each compound were first measured ( Figure S8). The maximum of the PLE spectra was used as the excitation wavelength for the PL emission measurements. In the first instance, the photoluminescence spectra of the coformers and pure ACA were measured, as depicted in Figure 5 and Figure S9. The emission maxima were observed at 509 nm (λ ex = 420 nm) for ACA, 372 nm (λ ex = 340 nm) for BPEE, and 418 nm (λ ex = 355 nm) for BPE. Upon excitation with 450-nm radiation, the ACA-BPEE emission maximum was observed at 492 nm slightly blue-shifted as compared with ACA ( Figure 5). The blue shifts of the emission maxima can be attributed to the H-type aggregation and formation of a robust hydrogen-bonded network in the cocrystal. In the case of the ACA-BPE cocrystal, the emission maximum showed a slight red shift to 518 nm (λ ex = 420 nm) as compared to ACA, presumably due to the absence of conjugation in the BPE molecule ( Figure 5). The perceived color of the coformers and the multicomponent solids are displayed on a CIE 1931 diagram (Commission internationale de l'éclairage, Figure S10).

Luminescent Property
Altering the crystallographic packing and intermolecular interactions of the organic luminophore can greatly influence the solid-state emission properties of a material. ACA itself is a well-known luminophore, and its luminescent properties were studied in the solution, as well as in the solid-state [54,55]. ACA therefore acts as an excellent candidate to study the effects of cocrystallization on luminescent properties.
To collect the photoluminescence spectra (PL), the PL excitation (PLE) spectra for each compound were first measured ( Figure S8). The maximum of the PLE spectra was used as the excitation wavelength for the PL emission measurements. In the first instance, the photoluminescence spectra of the coformers and pure ACA were measured, as depicted in Figure 5 and Figure S9. The emission maxima were observed at 509 nm ( ex = 420 nm) for ACA, 372 nm ( ex = 340 nm) for BPEE, and 418 nm ( ex = 355 nm) for BPE. Upon excitation with 450-nm radiation, the ACA-BPEE emission maximum was observed at 492 nm slightly blue-shifted as compared with ACA ( Figure 5). The blue shifts of the emission maxima can be attributed to the H-type aggregation and formation of a robust hydrogen-bonded network in the cocrystal. In the case of the ACA-BPE cocrystal, the emission maximum showed a slight red shift to 518 nm ( ex = 420 nm) as compared to ACA, presumably due to the absence of conjugation in the BPE molecule ( Figure 5). The perceived color of the coformers and the multicomponent solids are displayed on a CIE 1931 diagram (Commission internationale de l'éclairage, Figure S10). To better understand the effect of the coformers and interactions present in the cocrystals on the PL spectra, the electronic structure of each cocrystal was explored within the remit of the DFT. The electronic density of states (DoS) was calculated at the HSE06 level (Figure 6), which is known to reproduce the experimental band gaps well. [56] We note that the ACA excitation in our systems To better understand the effect of the coformers and interactions present in the cocrystals on the PL spectra, the electronic structure of each cocrystal was explored within the remit of the DFT. The electronic density of states (DoS) was calculated at the HSE06 level ( Figure 6), which is known to reproduce the experimental band gaps well [56]. We note that the ACA excitation in our systems exhibits a band gap of 435 nm (2.85 eV). Although this is indicative of the fundamental (single-particle) band gap, optical band gaps in molecular crystals typically differ by only a fraction of an eV [57]. Our calculated fundamental gap is in agreement with the excitation energy observed experimentally for ACA (420 nm), thereby lending support to the validity of the calculations. For both cocrystals, the top of the valence band is composed entirely of ACA-based states. ACA-BPEE exhibits coformer-based states at the bottom of the conduction band, whereas there are mixed ACA and coformer states in the case of ACA-BPE. It follows that direct excitation across the band gap in the former is associated with a charge transfer between the ACA and coformer molecules. The PL emission energies cannot occur from relaxation of the single electron states described in Figure 6. To explore this further, the asymmetric unit of the ACA coformer was extracted from the unit cell and modeled using time-dependent DFT calculations.
For each dimer in their crystallographic geometry, the vertical excitation energies were calculated in agreement with the experimental absorption values at the B3LYP/def2-TZVP level of the theory. Moreover, the predicted dimer excitation energy is in agreement with the simulated solidstate band gap and band character in Figure 6, thereby suggesting that the extracted dimers provide a good approximation for the solid-state electronic structure. In each case, the excitation corresponds to a direct excitation of an ACA-based orbital to a conformer-based orbital (Figure 7). This lends support to charge transfer-based excitation in the newly formed cocrystal phases. The PL emission energies cannot occur from relaxation of the single electron states described in Figure 6. To explore this further, the asymmetric unit of the ACA coformer was extracted from the unit cell and modeled using time-dependent DFT calculations.
For each dimer in their crystallographic geometry, the vertical excitation energies were calculated in agreement with the experimental absorption values at the B3LYP/def2-TZVP level of the theory. Moreover, the predicted dimer excitation energy is in agreement with the simulated solid-state band gap and band character in Figure 6, thereby suggesting that the extracted dimers provide a good approximation for the solid-state electronic structure. In each case, the excitation corresponds to a direct excitation of an ACA-based orbital to a conformer-based orbital (Figure 7). This lends support to charge transfer-based excitation in the newly formed cocrystal phases. calculated in agreement with the experimental absorption values at the B3LYP/def2-TZVP level of the theory. Moreover, the predicted dimer excitation energy is in agreement with the simulated solidstate band gap and band character in Figure 6, thereby suggesting that the extracted dimers provide a good approximation for the solid-state electronic structure. In each case, the excitation corresponds to a direct excitation of an ACA-based orbital to a conformer-based orbital (Figure 7). This lends support to charge transfer-based excitation in the newly formed cocrystal phases.

Conclusion
We herein report two new two-fold interpenetrated isostructural cocrystals of ACA with different dipyridines. Each new cocrystal was prepared by both the mechanochemical and solution methods. The new solids were characterized by SCXRD, thermal analysis (TGA and DTA), PXRD, FTIR, and Raman spectroscopy. Both the cocrystals are sustained by pyridine-carboxylic acid synthons via O-H· · · N hydrogen bonding to form a three-component supramolecular assembly, and the C-H···O hydrogen bonds were very vital for guiding the three-dimensional supramolecular assembly. The solid-state emission spectra of these new multicomponent solids signified that their luminescent properties significantly depended on the chemical nature of the coformers. The DFT calculations suggest that these modified luminescent properties are the result of a photo-excited charge transfer between the ACA and conformer molecules. We, therefore, demonstrate how cocrystallization can offer a powerful method to tune the emissions of organic lumiphores and suggest this to be an excellent approach for the design of various new types of organic luminescent materials.