Use of additives to regulate solute aggregation and direct conformational polymorph nucleation of pimelic acid

The interference of solute self-assembly caused by the interactions between pimelic acid and a series of homologous additives is closely related to the ability to induce the form II compound with similar packing but a different conformation to that of form I. The novel use of additives demonstrates the direct link between solute aggregation in solution and molecular conformation in crystals.

The background, the corresponding solvent to liquid samples was deducted in collecting spectrograms of samples. NMR spectra were detected using a 600 MHz JEOL JNM-ECZ600R/S1.

S1.3. Solubility measurement of DA7 with additives in 1, 4-dioxane
A gravimetric method was employed to determine the solubility of form I of DA7 with or without DA3/5/9/11 in 1, 4-dioxane (0.1464 mol additive / mol solvent, equal to the molar ratio 5:100 of additive to DA7 in above experiments) at 298.15 K. In this process, additive solutions in 1, 4-dioxane at the concentration of 0.1464 mol additive / mol solvent were prepared firstly. Then excess form I of DA7 and corresponding additive solutions, were added to 50 mL flasks so that to obtain the suspensions. Then the suspensions were shaken by a thermostatic bath shaker (CHY1015, Shanghai Sunny Hengping Scientific Instrument Co. Ltd., China) at a certain temperature under uncertainty of 0.1 K. And this process would last for 12 h which had been proved to be long enough to achieve solid-liquid equilibrium in preliminary experiment. After turning off the bath shaker, 5 mL of the supernatant liquor was filtered by the pre-cooled/heated syringes filters (0.22 µm) and moved into pre-weighted glass dishes as quickly as possible. Immediately, the total weight was determined. After that, the dishes were dried in a vacuum oven (DZ-2BC, Tianjin Taisite Instrument Co. Ltd., China) at T=343.15 K and their mass was periodically measured until the data remained constant, which meant that the solvent had been completely evaporated. In all above experiments, the masses were determined by an electronic balance (AB204-N, Mettler-Toledo, Switzerland) with an accuracy of ± 0.0001 g. The experiment was repeated three times for error reduction, and the result was from the average value.
The mole fraction solubility of DA7 (x 1 ) was calculated by using Eq. (S1-2): Where m s represents the mass of solid, m 1 represents the mass of solute DA7, m 2 mean the masses of analyzed by powder X-ray diffraction (PXRD) to identify the form ( Figure S1), was separated as soon as possible after nucleation during cooling to avert the underlying polymorphic transition (the transformation of pure form II in solution occur after 42 min since nucleation at least and that of mixtures of form I and II occur earlier (about 25 min since nucleation with high fraction of form I)).
The samples used to observe crystal shapes were taken out after 20 min since nucleation before phase transition.
Cooling experiments were conducted with the initial concentration of DA9 in dioxane of 0.439 g/ml (0.007 mol DA9 / 3 ml dioxane) with adding 9 additives in different additive concentrations (the molar ratio 1: 10 of additive to DA7). A given mass of DA7 form I and additives were added into 3 ml solvent and heated to 333.15 K to ensure solid was dissolved completely. Then the solutions were filtered through a preheated 0.22 μm syringe filter, and transferred into a jacketed vessel and held at 333.15 K for 30 min. After that the solution systems were cooled to 288.15 K with a cooling rate of 0.2 K/min while stirring (400 rpm) by a magnetic stirrer. A blank experiment was carried out without additives at the same condition (stable form I obtained). The obtained solid, analyzed by powder Xray diffraction (PXRD) to identify the form, was separated as soon as possible after nucleation during cooling to avert the underlying polymorphic transition.

S1.4.2. Isothermal experiments
To eliminate the influence of supersaturation and temperature in previous experiments, two groups of isothermal crystallization experiments at T=298.15 K with or without additives were carried out. In Group 1, the supersaturation of each batch was fixed at 1.5 based on thermodynamic data. In Group 2, we kept the absolute concentration of DA7 equivalent to the amount when S=1.5 without additives so that solutes and additives are in the same molecular proportion. Based on this, given masses of form I of DA7 and corresponding additives (DA3/5/9/11, 0.1464 mol additive/mol solvent) were added into 10 g pure 1, 4-dioxane based on the desired concentration at 298.15 K, and heated to elevated temperature (348.15 K) to make the solute dissolved completely. Then the heated solutions were filtered through a preheated 0.22 μm syringe filter and simultaneously transferred into a glass test tube and held at 298.15 K for 30 min. After that the tubes were rapidly transferred into a thermostatic water bath (model 501 A, Shanghai Laboratory Instrument Works Co., Ltd., China) with an accuracy of ± 0.05 K while agitating (400 rpm) by a magnetic stirrer. After the induction stage of nucleation, the solid appeared. Then they would be separated from the suspension as soon as possible and analyzed by PXRD to determine the crystal form. Each experiment was repeated three times.

S1.5. Calculation of electrostatic potential (ESP) charges
In this work, the ESP maps and ESP charges of additives were calculated at DFT level in DMol3 module using the molecular modeling software package Materials Studio (Accelrys Inc., USA).
Before analysis of ESP charges, the geometry optimization of molecules were performed at first by double numerical plus polarization (DNP) basic set. The Perdew-burke-ernzerhof (PBE) and generalized gradient approximation (GGA) were selected. The quality of self-consistent-field (SCF) tolerance was fine. The ESP charges were assigned on molecules based on analysis function in Dmol3 module. The calculations were carried out for two times for comparison: the structures of all molecules were generated manually or picked from crystal structures. It was shown that the results were little to do with the molecular source.

S1.6. Calculation of binding energies of DA7 with additives
Density functional theory (DFT) calculations have been applied using a Gaussian 09 package to investigate interactions in (1:1) molecular complexes of DA7 with additives in SMD solvent model (dioxane). Similarly with the Reference you provided, the equilibrium geometries of the modeled associates are calculated with a B97D Grimme's functional, which includes a long-range dispersion correction. This allows for better description of the van der Waals interactions and gives proper geometries of molecular clusters. Binding energies of the complexes are calculated using a double hybrid B2PLYPD functional, which combines exact Hartree−Fock exchange with an MP2-like correlation and long-range dispersion corrections. A Gaussian-type 6-31G (d,p) basis set is used for geometry optimization and a triple-ζ valence quality (TZVP) basis set is used for energy calculations

S4. Supplementary information for FTIR spectrum analysis
As shown in Figure S12, only the peak at about 1736 cm -1 is retained when DA7 in 1, 4-dioxane at low concentration, which should be the vibration band of non-hydrogen-bonded C=O groups due to the solvent with no hydrogen-bonding donors. With the concentration of DA7 increasing, a new band at about 1712 cm -1 appeared. And the higher the concentration, the higher intensity ratio of the band at 1712 cm -1 to that at 1736 cm -1 . It declares that the carboxyl groups of diacids self-assembly happens in 1, 4-dioxane after reaching to a certain concentration, and the degree of aggregation increases with the increase of concentration. It indicates that is intermolecular interaction of solutes rather than intramolecular interaction, e.g. molecule self-cyclization. It is of great difference from solvents with high HBD ability where the molecules exist as solvated or non-solvated monomers.

Figure S13
Solution IR spectra of pure DA7 over a concentration range in 1, 4-dioxane.
As shown in Figure S13, for DA2 and DA3 at both concentrations, there exist two vibration peaks of C=O, whose intensity ratio keep constant as concentration increases, respectively corresponding to free monomers (high wavenumber) and solvated monomers (low wavenumber). As additives, they also have a certain embodiment.
At the concentration of 0.25 mol solute/L solvent, there is no obvious signal of C=O aggregation for all diacids. When they work as additives, the peak intensity of solute aggregation at about 1712 cm -1 increases in all samples studied.
At the concentration of 0.6 mol solute/L solvent, there exists slight aggregation in DA6-11. However, when they work as additives, the peak intensity of solute aggregation increase obviously. More