Impact of Supramolecular Aggregation on the ... - ACS Publications

May 9, 2017 - scanning calorimeter (DSC) (TA Instruments, New Castle, ..... Similar peak splitting is seen for compounds 3 (Figure. 2D(i)) and 7 (Figu...
0 downloads 0 Views 2MB Size
Subscriber access provided by CORNELL UNIVERSITY LIBRARY

Article

Impact of Supramolecular Aggregation on the Crystallization Kinetics of Organic Compounds from the Supercooled Liquid State Arjun Kalra, Patrick Tishmack, Joseph W. Lubach, Eric Munson, Lynne S. Taylor, Stephen R. Byrn, and Tonglei Li Mol. Pharmaceutics, Just Accepted Manuscript • Publication Date (Web): 09 May 2017 Downloaded from http://pubs.acs.org on May 10, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Molecular Pharmaceutics is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

TOC Graphics

Synopsis: Supramolecular structures formed by organic molecules in the melt dictate subsequent kinetics of molecular packing into crystal. A hydrogen-bonded chain motif favored in the melt could face greater energy barriers compared with a dimer motif.

ACS Paragon Plus Environment

1

Molecular Pharmaceutics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 40

Impact of Supramolecular Aggregation on the Crystallization Kinetics of Organic Compounds from the Supercooled Liquid State Arjun Kalra,1 Patrick Tishmack,2 Joseph W Lubach,3 Eric Munson,4 Lynne S. Taylor,1 Stephen R. Byrn,1 and Tonglei Li 1* 1

Department of Industrial and Physical Pharmacy, College of Pharmacy, Purdue University, West Lafayette, Indiana 47907 2

AMRI SSCI, LLC, West Lafayette, Indiana 47906 3

4

Genentech, South San Francisco, CA 94080

Department of Pharmaceutical Sciences, College of Pharmacy, University of Kentucky, Lexington, KY 40508

* Corresponding Author: Tonglei Li, Ph.D. Department of Industrial & Physical Pharmacy College of Pharmacy Purdue University RHPH Building, Room 124B 575 Stadium Mall Drive West Lafayette, IN 47907-2091 E-mail: [email protected] Phone: (765) 494-1451

ACS Paragon Plus Environment

2

Page 3 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

ABSTRACT Despite numerous challenges in their theoretical description and practical implementation, amorphous drugs are of growing importance to the pharmaceutical industry. One such challenge is to gain molecular level understanding of the propensity of a molecule to form and remain as a glassy solid. In this study, a series of structurally similar diarylamine compounds was examined to elucidate the role of supramolecular aggregation on crystallization kinetics from supercooled liquid state. The structural similarity of the compounds makes it easier to isolate the molecular features that affect crystallization kinetics and glass forming ability of these compounds. To examine the role of hydrogen-bonded aggregation and motifs on crystallization kinetics, a combination of thermal and spectroscopic techniques was employed. Using variable temperature FT-IR, Raman and solid-state NMR spectroscopies, the presence of hydrogen bonding in the melt and glassy state was examined and correlated with observed phase transition behaviors. Spectroscopic results revealed that the formation of hydrogen-bonded aggregates involving carboxylic acid and pyridine nitrogen (acid-pyridine aggregates) between neighboring molecules in the melt state impedes crystallization, while the presence of carboxylic acid dimers (acid-acid dimers) in the melt favors crystallization. This study suggests that glass formation of small molecules is influenced by the type of intermolecular interactions present in the melt state and the kinetics associated with the molecules to assemble into a crystalline lattice. For the compounds that form acid-pyridine aggregates, the formation of energy degenerate chains, produced due to conformational flexibility of the molecules, presents a kinetic barrier to crystallization. The poor crystallization tendency of these aggregates stems from the highly directional hydrogen-bonding interactions needed to form the acid-pyridine chains. Conversely,

ACS Paragon Plus Environment

3

Molecular Pharmaceutics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 40

for the compounds that form acid-acid dimers, the non-directional van der Waals forces needed to construct a nucleus promote rapid assembly and crystallization.

Keywords: glass-forming ability, crystallization, nucleation, amorphous, spectroscopy, molecular interactions

ACS Paragon Plus Environment

4

Page 5 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

INTRODUCTION Amorphous formulations of drug molecules are of growing importance to the pharmaceutical industry.1-4 The low aqueous solubility and consequent poor bioavailability associated with the crystalline state of many newly developed drug candidates has propelled amorphous formulations into greater relevance.5-6 However, this high-energy state possesses excess thermodynamic quantities including free energy, enthalpy and volume, which gradually drift towards equilibrium values leading to either nucleation and subsequent crystallization or, in other cases, structural relaxation (annealing) to a lower energy glassy state.7-9 Understanding and predicting the stability of amorphous solids remains one of the key questions for successful application of this solid state in formulation design. Crystallization kinetics of organic drug molecules from supercooled melts has attracted considerable research attention.10-17 The propensity of a material to vitrify is referred to its glass forming ability (GFA).18-19 Recent work of seeking molecular understanding of glass formation has highlighted some of the key molecular descriptors impacting the vitrification of organic molecules. Properties such as molecular branching, a high number of rotatable bonds, and distribution of electronegative atoms in molecules were positively associated with glass formation, while other properties such as low molecular weight and high number of benzene rings were found to promote crystallization.10, 12-13 Because crystallization involves molecular assembly through intermolecular interactions, the role of hydrogen bonding on glass formation from melts has been studied using spectroscopic and computational methods.20-22 For 2biphenylmethanol, the formation of hydrogen-bonded tetramers during the vitrification process has been reported and correlated with glass formation.23 A group of structurally similar triazine compounds was examined and it was inferred that strong intermolecular interactions such as

ACS Paragon Plus Environment

5

Molecular Pharmaceutics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 40

hydrogen bonding could promote amorphization by forming aggregates that pack poorly.24-29 Combining experimental studies with computation, one study speculated that the presence of two distinct types of hydrogen-bonded motifs, which pack poorly, was responsible for the good glass forming ability of salsalate.22 Conformational influence on crystallization kinetics was studied for various systems including sugar alcohols.30-31 It was found that sorbitol and iditol retain their glassy state upon cooling from the melt, while the structurally related mannitol and dulcitol crystallize under similar conditions.31-33 It was thus suggested that hindered crystallization of sorbitol or iditol results from the presence of distinct conformers in the melt and the solid state.30, 33

Similarly, it was suggested that the existence of multiple molecular conformers of 2-methyl-

1,3-propanediol prevents crystallization, while 1,3-propanediol, structurally similar but lacking the methyl group, crystallizes readily.34 It is apparent that certain molecular features play an important role in glass formation. However, understanding remains limited with respect to the role that these molecular features play in the formation of supramolecular assemblies during the initial stage of nucleation and subsequent crystallization or amorphization. In this study, we compare the vitrification potential of a series of structurally similar diarylamine compounds by analyzing molecular assemblies in the melt, glassy and crystalline states. The diarylamine scaffold was systematically modified to understand key molecular features on crystallization kinetics and glass-forming ability. The set of compounds is divided into two broad categories based on their chemistry (Scheme 1). Type 1 compounds (1-6) have both carboxylic acid and pyridine functional groups and can form both acid-acid and acid-pyridine hydrogen-bonded motifs. In addition, they possess an intramolecular hydrogen bond between the bridging amino group and the carboxylic acid in the crystal structures (except compound 6). Type 2 compounds (7-9) lack the pyridine N and thus cannot

ACS Paragon Plus Environment

6

Page 7 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

form acid-pyridine hydrogen bonding; the intramolecular hydrogen bond still exists in the crystal structures.

ACS Paragon Plus Environment

7

Molecular Pharmaceutics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 40

EXPERIMENTAL SECTION Materials Based on previously established methods, compounds 1, 2, 3, 4, and 6 were synthesized and purified in our laboratory and the chemical structures and purity were verified using solution NMR spectroscopy.35-38 Compounds 5, 7, 8 and 9 were purchased from TCI America (Portland, OR) and used as provided (purity>98%). Solvents (acetonitrile, methanol, dichloromethane, hexane, and ethyl acetate; ACS grade) were purchased from Macron (Center Valley, PA) and ethanol (200 proof; USP grade) was obtained from Decon Labs (King of Prussia, PA). Methods X-ray powder diffraction (XRPD): To verify the polymorphic forms of the compounds, a Rigaku Smartlab diffractometer (The Woodlands, TX) with Cu Kα radiation (40kV, 40mA) was used to collect diffraction patterns between 4.0° and 40.0° 2θ in the Bragg-Brentano configuration. A step size of 0.02° with a scan rate of 4°/min was used for data collection at room temperature. Differential scanning calorimetry: A Q20 differential scanning calorimeter (DSC) (TA Instruments, New Castle, DE) equipped with a liquid nitrogen cooling system (LNCS) was utilized for characterizing phase transition behaviors of the compounds. Temperature and heat flow were calibrated by measuring the onset temperature and enthalpy response of an indium standard. Tzero® hermetic aluminum pans and lids were used in the experiments with a constant purge of helium gas at 25 mL/min. Samples were hermetically sealed and heated at 10 °C/min to approximately 3-5 °C above the melting offset. The samples were then cooled at the rate of 20 °C/min to −40 °C, and then reheated again above the melting temperature at 10 °C/min.

ACS Paragon Plus Environment

8

Page 9 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

Variable-temperature FT-infrared spectroscopy: A Bruker Vertex 70 FTIR spectrometer (KBr beam splitter, DTGS detector and 6 mm aperture) fitted with a Golden-Gate diamond attenuated total reflectance (ATR) temperature-controlled accessory (Specac, Fort Washington, PA) was used to collect spectra at various temperatures. A total of 64 scans at a resolution of 4 cm−1 was obtained. Because the IR transmission of diamond shows temperature dependence, background spectra were also collected at each sampling temperature. OPUS (Bruker Optic, Version 7.2) was used to process the collected spectra. Raman spectroscopy: A Bruker Senterra Raman microscope equipped with a 785 nm laser and 20× objective was used to collect sample spectra at a spectral resolution of 4 cm−1. A total of 20 co-additions with an integration time of 5 sec was used for each measurement. OPUS (Bruker Optic, Version 7.2) was used to process the Raman spectra. Preparation of amorphous (glassy) samples for SSNMR experiments: Amorphous (glassy) samples were prepared by melt quenching. A known quantity of a crystalline sample was heated a few degrees above its melting point, and the melt was then immediately immersed in liquid nitrogen (−196 °C). Vitrification of the sample was verified using XRPD and DSC. The amorphous samples for solid-state NMR experiments were prepared immediately before data collection. However, if storage was required, the samples were stored in a refrigerator at approximately 4 °C. 13

C solid-state NMR: An Agilent DD2-400 MHz (1H = 399.82MHz;

13

C = 100.55 MHz)

SSNMR spectrometer was used to obtain spectra of crystalline/amorphous compounds 1 and 2. Samples were packed in 4 mm PENCIL type zirconia rotors and

13

C SSNMR spectra were

obtained with cross-polarization/magic angle spinning (CP/MAS) and total sideband suppression (TOSS) at ambient temperature. Additional operating conditions included a spinning speed of 12

ACS Paragon Plus Environment

9

Molecular Pharmaceutics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 40

kHz, acquisition time of 30 ms, phase-modulated (SPINAL-64) high-power proton decoupling during acquisition, and an 1H pulse width of 2.9 µs (90°). A CP contact time of 5 or 2 ms was used for crystalline or amorphous samples, respectively. A Chemagnetic CMX 400 MHz (1H = 400.02MHz; 13C = 100.59 MHz) spectrometer was used to obtain the spectra of compounds 3, 4, 5, 7 and 8. Samples were packed in a 5mm zirconia rotor and 13C spectra were obtained with cross-polarization/magic angle spinning (CP/MAS) and total sideband suppression (TOSS) at ambient temperature. Operating conditions included a spinning speed of 6 kHz, acquisition time of 16 ms, TPPM high power proton decoupling, and an 1

H pulse width of 5 µs (90°). A CP contact time of 3.5 ms was used for all samples. The pulse delay time was different for each compound, with amorphous samples requiring

much less recycle delay than crystalline samples. The

13

C SSNMR spectra were externally

referenced to the carbonyl carbon resonance of glycine at 176.5 ppm. The free induction decays were processed using Agilent VnmrJ 3.2A or Spinsight Version 4.3.2 software and an exponential line broadening factor of 20 Hz was used to improve the signal-to-noise ratio. Ab initio calculations and molecular modeling: The hydrogen-bonding energy of the molecules was calculated by quantum mechanics.39 Experimentally determined crystal structures were optimized at the B3LYP/6-21G(d,p) level with the lattice parameters fixed by Crystal09.40 From the optimized crystal structures, hydrogen-bonded pairs were extracted and subjected to single-point energy calculations at B3LYP/6-311++G(d,p) by Gaussian09.39 The hydrogenbonding energy was calculated by subtracting twice the monomer energy from the energy of the dimer (acid-acid or acid-pyridine) with the basis set superposition error (BSSE) corrected by the counterpoise method.41

ACS Paragon Plus Environment

10

Page 11 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

For molecules that do not form crystal structures with the acid-acid or acid-pyridine synthon, hydrogen-bonding strength of the synthon was calculated by taking the energy minimum from a plot of hydrogen-bonding energy of the motif versus intermolecular O/O distance of the neighboring acid-acid group or the O/N distance of the hydrogen-bonded pair. The pair of molecules for each data point were partially optimized with the O/O or O/N distance fixed. Lattice energies of the α and δ polymorphs of 1 were calculated with an empirically augmented density functional theory (DFT) method.42 The crystal structures were firstly optimized with the lattice parameters held constant while allowing fractional coordinates of all atoms to adjust. The basis set used was B3LYP/6-21G(d,p); no diffuse orbitals were included due to potential instability introduced to Bloch functions. The single-point energy of optimized crystal structures was then calculated and the van der Waals energy component was included through empirical equations. The lattice energy was computed as the difference between the total corrected energy of a crystal system and the lowest energy of a fully optimized single molecule in vacuum. Hirshfeld surfaces for the acid-acid and acid-pyridne dimers of 1 were generated using Crystal Explorer (Version 3.1, Revision 1448).43 The optimized crystal structures of 1 were used for the calculation.

ACS Paragon Plus Environment

11

Molecular Pharmaceutics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 40

RESULTS Polymorph crystallization and crystal structure analysis To understand the phase behavior, especially with regard to preference for crystallization or amorphization, crystal structures of all compounds were analyzed, including molecular packing and the preferred hydrogen-bonding motif (Figure 1). Polymorphic forms obtained after crystallization were verified using powder X-ray diffraction. Compound 1 is known to have four polymorphs (α, β, γ, δ).35 In the α polymorph (Figure 1a), the molecules adopt a –COOH···–COOH [HOC=O···HOCO] hydrogen-bonded dimer synthon while in the δ polymorph, –COOH···pyridine N hydrogen-bonded catemers (Figure 1b) are observed. A total of nine different conformers, mainly differing in torsion angle τ1, are found in all four crystal structures with the δ polymorph possessing two crystallographic independent molecules in its asymmetric unit.35 Compound 2 (Figure 1c) has one reported polymorph which adopts the acid-acid homosynthon.37 In order to discover possible polymorphs with acid-pyridine chains, polymorph screening was attempted. Despite the effort, no new form was obtained. Compound 3 displays one-dimensional chains of the acid-pyridine heterosynthon in the reported crystal structure (Figure 1d).37 The crystal contains 8 molecules in the unit cell with Z’=2.37 Polymorphic screening in various solvents yielded only the reported solid form, and no acid-acid homosynthon containing crystal structure was found. Molecules in compound 4 pack as acidacid dimers with Z’=1 (Figure 1e).38 A new polymorphic form of this compound was obtained during crystallization from the amorphous sample and will be discussed elsewhere. Compound 5 (clonixin) has four reported polymorphs (I, II, III and IV).44 While forms I, III and IV are neutral, polymorph II is zwitterionic. Form I is the most stable form at room temperature and is enantiotropically related to forms III and IV.45 In polymorph I, the molecules adopt –

ACS Paragon Plus Environment

12

Page 13 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

COOH···pyridine N hydrogen-bonded catemers (Figure 1f), while in the polymorphs III (τ1= 160°) and IV (τ1=178°) –COOH···–COOH hydrogen-bonded dimers are observed (Figures 1g and 1h respectively). Compound 6 has four reported anhydrous polymorphs (I, II, III, IV) and one monohydrate form.36 All forms possess acid-pyridine hydrogen bonded catemers. Form I, the most stable form, was used for the subsequent experiments.36 Compound 7 (fenamic acid) lacks the pyridyl N and adopts the acid-acid dimer synthon (Figure 1i). There are two different conformers (Z’=2) in the asymmetric unit of the only crystal structure that is reported in the literature.46 Compound 8 (mefenamic acid) has three reported polymorphs (I, II and III). At room temperature, Form I is the most stable form and was provided by the supplier.47 These conformational polymorphic forms exhibit –COOH···–COOH hydrogen-bonded dimers in the crystal structures. Compound 9 (tolfenamic acid) has five reported crystal structures but only two polymorphs are commonly encountered.48 Form I (colorless) and Form II (yellow) II differ mainly in the torsion angle τ1 (−74.9° and −142.6° of forms I and II, respectively) with similar – COOH···–COOH hydrogen-bonded dimers in the two forms (Figure 1j). Form I is reported as the most stable polymorphic form and was provided by the supplier.49 The crystal structure analyses demonstrate that these molecules can adopt either acid-acid or acid-pyridine synthon in the crystal structure suggesting that, if the molecules have to crystallize from the supercooled liquid upon cooling and form a glassy solid, they have to either form – COOH···pyridine N catemers or –COOH···–COOH dimers during the early stages of crystallization. DSC evaluation of crystallization from supercooled liquid state Phase transition kinetics of the compounds along with thermal properties (melting points,

ACS Paragon Plus Environment

13

Molecular Pharmaceutics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 40

enthalpies of melting, and glass transition temperatures) were determined by DSC (Table 1). On cooling from the melt state, compounds 1, 2, 3 and 6 avoided crystallization and formed glassy solids. On the other hand, compounds 4, 5, 7, 8 and 9 showed exothermic crystallization peaks in the supercooled melt region; respective onset crystallization temperatures were 124.0, 164.5 156.2, 180.6, and 143.9 °C and the recrystallized samples were identified by XPRD as the same polymorphs as the starting materials. The inability of compound 8 (mefenamic acid) and compound 9 (tolfenamic acid) to form glasses was reported previously (using melt quenching, spray drying and/or cryomilling).10,

50

On reheating, compounds 1 and 2 recrystallized and

subsequently demonstrated melting peaks. For compound 2, the melting point of the recrystallized solid was identical to that of the starting material; however, for compound 1, the recrystallized solid melted at the same temperature as the β polymorph. Phase transitions, such as crystallization, can be affected by the cooling rate employed during experiment, with rapid cooling rates often precluding re-crystallization events.10, 51-52 Our DSC experiments utilized a moderate cooling rate of 20 °C/min. To examine the effect of higher cooling rates, the melts were quenched with liquid nitrogen. Under these conditions, all compounds demonstrated the same phase behavior except compounds 4 and 5, which turned into glassy solid with Tg’s of 38.6 °C and 51.7 °C respectively. Upon reheating, 4 recrystallized into a new polymorph (as indicated by SSNMR and discussed later), which underwent solid-solid phase transition to the original form prior to melting; 5 recrystallized into Form IV and transformed to Form I (starting material). Based on their crystallization tendency, our compounds were divided into three groups based on the scheme reported by Baird et al.10 Class I compounds are rapid crystallizers (IA crystallize even under rapid cooling with liquid nitrogen or IB - vitrify upon cooling with liquid

ACS Paragon Plus Environment

14

Page 15 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

nitrogen). Class II compounds are good glass formers with poor stability on re-heating, and Class III compounds are good glass formers that exhibit stability on re-heating. From Table 1, it can be seen that compounds containing the pyridine N generally fall into Class II/III/IB and hence are glass formers. In contrast, compounds lacking this structural feature crystallize rapidly. This correlation is best exemplified by comparing compound 1 with 7 and compound 5 with 9, which differ only in the pyridine N, yet display distinct phase behaviors. Spectroscopic investigation of molecular interactions To understand the distinct phase behaviors of these structurally similar compounds, intermolecular interactions in the amorphous state were examined and compared to those found in the crystals.

13

C SSNMR spectroscopy is a powerful technique because of the sensitivity of

chemical shift to the electronic environment of a carbon atom in a solid sample. The resonant frequency of the carboxylic acid carbonyl carbon (typically observed between 165-175 ppm) is extremely sensitive to the hydrogen-bonding motif.53 As seen in Table 2, 13C SSNMR chemical shifts of crystalline solids containing the acid-acid synthon are shifted to higher frequency by 3-5 ppm compared with the acid-pyridine synthon. The carbonyl carbon is more deshielded in the acid-acid synthon, leading to the upfield shift. This is because in the acid-pyridine motif, the carbonyl carbon bears more electron densities (shielded) thanks to the significantly nucleophilic nitrogen from the pyridine ring. In fact, our calculation of partial atomic charges indicate the carbon in the acid-acid hydrogen bonding is more positive (deshielded) than the carbon in the acid-pyridine. As discussed later, the acid-pyridine hydrogen bonding is much stronger than the acid-acid.

ACS Paragon Plus Environment

15

Molecular Pharmaceutics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 40

13

C SSNMR spectra of crystalline (α, δ polymorph) and amorphous samples of compound 1

are shown in Figure 2A. The spectrum of the δ polymorph (Figure 2A(ii)) displays peak splitting for most carbon atoms. This is characteristic of two conformationally distinct molecules in the asymmetric unit (Z’=2). Similar peak splitting is seen for compound 3 (Figure 2D(i)) and compound 7 (Figure 2F) where Z’=2. Of particular significance is the resonance frequency of the carboxylic acid carbon because of its involvement in formation of hydrogen bonding (Table 2). For the α polymorph (Figure 2A(i)), δ polymorph (Figure 2A(ii)) and amorphous form (Figure 2A(iii)), the chemical shifts are observed at 174.4, 169.1, and 169.9 ppm, respectively. Similar values in the melt quenched amorphous sample and the δ form suggest that the electronic environment of carbonyl carbon is similar, corresponding to the carboxylic acid-pyridine intermolecular interaction. Analogous results were observed for other glass forming compounds. For compound 5 (shown in Figure 2B), the chemical shift of the carbonyl carbon for acidpyridine synthon containing Form I (Figure 2B(i)) and amorphous phase (Figure 2B(iv)) occur at 169.1 and 169.5 ppm, respectively. For Form III (Figure 2B(ii)) and IV (Figure 2B(iii)), the chemical shifts are 174.9 and 172.7 ppm, respectively. For compound 3, a peak position of 169.0-170.4 ppm (Z’=2) was detected for Form I (containing the acid-pyridine synthon) and 169.6 ppm for amorphous samples. In the case of compound 2, the glassy sample was prepared by melt quenching and the carbonyl carbon peak appeared at 169.6 ppm, compared with 174.0 ppm for the crystalline sample (acid-acid synthon). A polymorph containing the acid-pyridine synthon could not be obtained by polymorph screening for this compound. Finally, while compound 4 could be converted to the glassy phase by melt quenching with liquid nitrogen, it displayed rapid crystallization. For Form I (Figure 2E(i)), the carbonyl carbon peak appears at 174.0 ppm, while for the partially crystallized sample (Figure 2E(iii)), two distinct peaks are

ACS Paragon Plus Environment

16

Page 17 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

seen at 174.0 and 170.1 ppm. The new polymorph (Form II) shows the carbonyl carbon peak at 174.3 ppm, thus indicating the presence of the acid-acid synthon. For compounds 7, 8 and 9, NMR analysis was not possible because of rapid crystallization and inability to prepare amorphous samples by melt quenching. These results indicate that the compounds bearing the acid-pyridine interaction in the amorphous state is more difficult to crystallize. It should be noted that the amorphous spectra are generally much more broadened than those of crystalline samples, reflecting a larger range of molecular environments including conformation variations and interaction diversities. The results can only point out the domination of the acid-pyridine interaction. To further investigate the intermolecular interactions, mainly hydrogen bonding, in the amorphous state, variable temperature FTIR and Raman spectra were collected. The carbonyl and pyridine functional groups in our compounds can be either involved in –COOH···pyridine N or –COOH···–COOH interactions. The carbonyl stretching vibration thus provides a useful way of identifying molecular speciation by comparing the amorphous/melt spectra with those of the crystalline samples. Examining the melt state is especially critical for compounds 7, 8 and 9 since these could not be prepared as amorphous samples for SSNMR analysis. Carbonyl stretching for conjugated carboxylic acids is usually detected in the region of 16501800 cm−1, with strong IR bands, while Raman bands are often of weak or medium intensity.54 FTIR spectra of various states of compound 1 are shown in Figure 3A, including α polymorph, δ polymorph, melt, and amorphous solid. For the α polymorph (Figure 3Ai), the peak at 1648 cm−1 with a shoulder at 1660 cm−1 denotes the overlap of the in-plane deformation of N–H peak with the asymmetric carbonyl symmetric stretching vibration, while for the δ polymorph (Figure 3Aii), these vibrational bands are observed overlapping with peaks roughly at 1643 and 1663

ACS Paragon Plus Environment

17

Molecular Pharmaceutics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 40

cm−1. The in-plane deformation of N–H in the region 1650–1600 cm−1 that overlaps with the C=O stretching band has been observed for similar compounds.55-56 Furthermore, the lower wavenumbers associated with these bands can be explained by the presence of intramolecular hydrogen bonding between the bridging N-H and C=O. This is verified by comparing the FT-IR spectrum of compound 6 (Figure 3B) that shows the carbonyl stretching vibration at 1712 cm−1 because the N-CH3 group prevents formation of the intramolecular hydrogen bonding. In the melt state of compound 1, the two overlapping bands appear as distinct peaks at 1643 and 1684 cm−1. The peak at 1643 cm−1 is again assigned to the in-plane deformation of N–H, while the other peak at 1684 cm−1 is assigned to carbonyl stretching vibration. This peak position shifts to 1679 cm−1 in the amorphous solid sample. The vibrational band positions for compounds 2, 3 and 4 are reported in Table 3. Unlike compounds 1-6, compound 7 cannot form acid-pyridine but only forms acid-acid hydrogen bonding because it lacks the pyridine nitrogen (Scheme 1b). Figure 3C shows the carbonyl stretch region of the sole reported crystalline sample with the peak position at 1652 cm−1 measured at room temperature. As the solid melts, the peak shifts from 1652 to 1656 cm−1. This suggests that even on melting, acid-acid hydrogen-bonded dimers are retained in the melt although the interaction may be weakened. Similar results were obtained for compounds 8 (Figure 3D) and 9 (Table 3). The Raman spectra (Figure 4) were examined for the corresponding symmetric C=O stretch in the region of 1650-1690 cm−1. While the FTIR spectrum of compound 1 shows the asymmetric stretch, the symmetric stretching is inactive for the α polymorph in Raman spectrum (Figure 4a). However, the δ polymorph shows a weak symmetric stretch (in addition to the asymmetric stretch in IR spectra) band at 1660 cm−1 (Figure 4b). The Raman spectrum of the amorphous

ACS Paragon Plus Environment

18

Page 19 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

sample shows a weak band at 1682 cm−1 (Figure 4b). The amorphous state shows both IR and Raman stretch bands similar to those of the δ polymorph. Combined with the SSNMR spectroscopic data, it was concluded that the melt and amorphous states of compound 1 contain acid-pyridine hydrogen bonded aggregates. Similar results were observed for compound 5 (clonixin) with the polymorphic Form I showing both IR and Raman peaks at 1677 cm−1 (Table S1 and Figure S1 in Supporting information). For Forms III and IV, only asymmetric stretching vibrations show IR active bands with no corresponding peaks in Raman spectra. For the amorphous phase, the IR and Raman peaks are observed at 1679 and 1680 cm−1, respectively. These results point to the absence of acid-acid dimers in the amorphous state, and combined with SSNMR data, suggest the existence of acid-pyridine hydrogen bonded aggregates. Clearly, for those compounds that can form both acid-acid and acid-pyridine hydrogen-bonding motifs, choosing to form the acid-pyridine motif in the melt/amorphous state is correlated with their glass forming abilities.

ACS Paragon Plus Environment

19

Molecular Pharmaceutics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 40

DISCUSSION Understanding glass-forming ability is of extreme importance in developing amorphous formulations for poorly water-soluble drugs. A fundamental understanding of molecular factors that influence the vitrification potential is critical. Studying the crystallization kinetics of a molecule from its liquid (melted) state upon cooling could shed light on its glass-forming ability. Crystallization kinetics is dictated by the energy barrier for the molecules in the melt/amorphous state to build up long-range, orderly packed supramolecular structures. The influence of molecular structure on crystallization kinetics is highlighted by the differing phase transition behaviors of structurally similar compounds in this study, which, by and large, cannot be predicted a priori.24, 57-58 In this work, the effect of molecular structure on crystallization from supercooled melts was explored by studying a series of structurally similar compounds. The results from this study show that the formation of –COOH···pyridine N interactions over –COOH···–COOH in the melt state contributes to glass formation and stability, as shown by compounds 1-6. The experimental results also suggest that the presence of –COOH···–COOH interactions in the melt state facilitates rapid crystallization on cooling. This was observed for compounds 7, 8 and 9, where the hydrogen-bonded acid-acid dimers nucleate and crystallize readily. This is evident from the spectroscopic data that probes the hydrogen bonding of the molecules in the melt, crystalline and amorphous samples of the compounds. Variable temperature FTIR was utilized to examine the hydrogen bonding state of compounds 7-9 in the melt state. Because these compounds form centro-symmetric carboxylic acid dimes in their crystal structures, the change in carbonyl stretching vibration was examined on melting. The small shift observed in this vibration band (Table 3) indicates that these molecules preserve the carboxylic acid dimer configuration, although this interaction is slightly weaker than that in the

ACS Paragon Plus Environment

20

Page 21 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

solid state. Furthermore, the aromatic ring vibrations (Table 3) in the melt state also show slight deviation from that in the crystalline state, indicating conformational changes of the molecules. 13

C SSNMR spectroscopy proved to be a valuable technique in distinguishing specific hydrogen-

bonding interactions in the amorphous solid state. The acid-pyridine and acid-acid synthons demonstrate a difference of 3-5 ppm in chemical shift of the carbonyl carbon. On average, for the acid-acid synthon-containing crystal structures, the chemical shift of the carbonyl carbon is 174.5 ± 1.2 ppm. On the other hand, for the acid-pyridine crystal structures and amorphous samples, the respective chemical shift values are 169.3 ± 0.3 and 169.7 ± 0.2 ppm. Comparing compounds 1 with 7 and 5 with 9, respectively, the role of intermolecular interactions on crystallization kinetics is clearly illustrated. The lack of pyridine N in compounds 7-9 renders the molecules poor glass formers. The formation of acid-pyridine aggregates over acid-acid dimers in the melt of compounds 1-3 was further examined by calculating the strength of the hydrogen-bonding interactions (Table 4). It can be seen that the acid-pyridine hydrogen bonding is stronger than that of carboxylic acid dimers of these compounds by approximately 20 kJ/mol (per hydrogen bond; note that one molecule can form two acid-pyridine or two acid-acid hydrogen bonds). This likely leads to formation of locally preferred molecular pairs and even large associations via the pyridine N···acid hydrogen bonding. Nonetheless, these hydrogenbonded chain motifs may need to convert to the acid-acid pairs in forming a crystal (as observed during re-crystallization of compounds 1, 2, 4 and 5 during the re-heating the amorphous sample in DSC experiments (data not shown)). This is because, collectively, a crystal with the acid-acid pair may have a stronger lattice energy than a crystal with acid-pyridine motif of the same molecule. For example, the α polymorph of compound 1 has stronger calculated lattice energy (−113.91 kJ/mol) than the δ form (−98.04 kJ/mol). The negativity of a lattice energy is of

ACS Paragon Plus Environment

21

Molecular Pharmaceutics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 40

attraction; the α polymorph contains the acid-acid synthon and the δ form the acid-pyridine. Other intermolecular interactions including π–π stacking and van der Waals forces become more significant when more molecules start to encase an acid-acid pair. Clearly, disrupting the hydrogen-bonded molecular species in the liquid state in order to form a new motif in the crystal results in kinetic obstacles. Furthermore, alteration of the hydrogen-bonding motifs requires the molecules to change their conformations. The acid-pyridine motif always associates with twisted conformations while the acid-acid with planar ones. As such, both inter- and intramolecular interactions need to be adjusted during the recrystallization from the supercooled liquid, contributing to the kinetic barrier of phase transition. Conversely, the kinetic process becomes relatively easy when the hydrogen-bonded pairs remain unchallenged from the melt state to the crystalline state. Apparently, compounds 7-9 bear this trait. When acid-pyridine molecular aggregates are present in the supercooled liquid, crystallization is impeded, thereby promoting vitrification. Frustration in liquid structure is considered an important factor in the inhibition of crystallization, in which a considerable difference in the locally favored structure in the melt state and crystalline state is considered to play a crucial role.59 While liquid/melt lacks long-range order, it can possess short-range ordering due to specific interactions between molecules (or atoms). According to Tanaka’s two-order-parameter description of liquids,60-61 in any liquid there exist two local orderings or structures competing with each other, i.e., normal liquid where molecules are randomly packed and locally favored structures where molecules may form some ordering via intermolecular interactions. These locally favored structures are believed to have a symmetry inconsistent with that of the corresponding crystal structure and present kinetic barriers for crystallization.60-61 For compounds 1-6, it is possible that acid-pyridine aggregates form multiple hydrogen-bonded

ACS Paragon Plus Environment

22

Page 23 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

catemers (locally favored structures), which are energy similar and thereby having little driving force for crystallization. Torsion angle τ1 (Scheme 1) of these molecules shows large conformational diversity but these large variations differ by only a few kJ/mol. If such a compound prefers the acid-acid motif in the crystal structure, realignment of molecular speciation during the nucleation process could present a formidable barrier for the self-assembly of crystallizing species. Conversely, if the locally favored dimer structures are also energyfavored in the crystal, crystallization could readily occur. Furthermore, low molecular mobility is associated with acid-pyridine hydrogen-bonded chains, which can contribute to the glass-forming ability. This is supported by viscosity measurements of melted samples (results to be disseminated separately). According to Classical Nucleation Theory (CNT), nucleation rate (J) can be described by an Arrhenius equation in which, the pre-exponential factor, ‫ ܣ‬, is determined by a term describing the frequency of attachment (f*) of building units to the growing cluster (Equations 1 and 2).62-63 ஻

J = A exp (- ௞ ்)

Equation 1

A = z f* C

Equation 2



where B is free-energy barrier for nucleus formation, kB is the Boltzmann factor, T is the temperature, z is Zeldovich factor, and C is nucleus concentration. f* is determined by the diffusion coefficient of growth units, which is influenced by energy barriers originating from conformational change associated with incorporation of molecules in a nucleus.62-63 The low addition frequency of acid-pyridine hydrogen bonded aggregates could be explained by the highly directional hydrogen bonding interaction needed to form catemers (Figure 5) as well as the conformational diversity in the melt/amorphous state. The changes in aromatic ring

ACS Paragon Plus Environment

23

Molecular Pharmaceutics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 40

vibrations between crystalline and amorphous states of compounds 1-6 shown in Table 3 indicate conformational diversity in the amorphous state (with respect to the crystalline state). Apparently, these locally formed chain structures lead to increase in the free-energy barrier for nucleation (B in Equation 1).64 In contrast, the non-directional van der Waals forces needed to construct a nucleus from carboxylic acid dimers promotes rapid assembly and crystallization (Figure 5). Such difference in molecular speciation in the melt state is believed to be the determining factor for the crystallization kinetics observed in this study.

ACS Paragon Plus Environment

24

Page 25 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

CONCLUSIONS This study examined the role of intermolecular interactions on the crystallization kinetics and glass forming ability of a series of structurally similar organic molecules. It was discovered that the acid-pyridine hydrogen-bonded association of molecules in the melt state presents a significant kinetic barrier to crystallization. On the other hand, the presence of carboxylic acid hydrogen-bonded dimers in the melt state resulted in rapid crystallization. The impeded crystallization kinetics of energy-similar acid-pyridine chains may be a result of poor efficiency in attachment of growth units, realignment of molecular conformations, and transformation of intermolecular hydrogen-bonding patterns. When dissimilar hydrogen-bonding synthons are formed respectively in the amorphous and crystalline states and, more importantly, revolution of the synthons during the phase transition requires molecules to vary their conformations, recrystallization from the amorphous state thus becomes kinetically frustrated. This work highlights the role of supramolecular aggregates in the melt state and, broadly, the interplay between molecular interactions and conformations to influence glass formation and recrystallization kinetics.

ACS Paragon Plus Environment

25

Molecular Pharmaceutics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 40

ACKNOWLEDGEMENTS The authors would like to thank Dr. Keith Chadwick (Purdue University) for valuable discussions. Financial support from PRF (Purdue Research Foundation) and CPPR (Center for Pharmaceutical Processing Research) is acknowledged.

ACS Paragon Plus Environment

26

Page 27 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

Scheme 1. Chemical structures of compounds examined in this study. Major torsional angles are highlighted in 1.

ACS Paragon Plus Environment

27

Molecular Pharmaceutics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 40

Figure 1. View of major synthons seen in the crystal structures of select compounds: 1 (α polymorph) – (a), Compound 1 (δ polymorph) – (b), Compound 2 – (c), Compound 3 – (d), Compound 4 – (e), Compound 5 (Form I) – (f), 5 (Form III) – (g), 5 (Form IV) – (h), Compound 7 – (i), Compound 9 (Form 1) – (j). Synthons for compounds 6 and 8 are not shown.

ACS Paragon Plus Environment

28

Page 29 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

Table 1. Physical and Thermal Properties of the Compounds Examined in the Study.

Compounds

MW (g/mol)

Tm onset (°C)

Enthalpy of melting (J/g)

Tg onset (°C)

Class

98.4

46.85

II

α: 153.26 1

214.22

β: 145.00 δ: 136.34

2

228.25

163.69

115.0

43.32

II

3

256.30

168.26

90.2

45.59

III

4

232.21

153.27

94.4

38.64*

IB

5

262.69

239.40

128.5

51.70*

IB

6

228.25

130.88

116.0

26.42

7

213.23

185.51

142.2

---

8

241.29

231.54

140.0

---

IA

9

261.71

211.98

137.1

---

IA

III

IA

* Measured after melt quenching with liquid nitrogen

ACS Paragon Plus Environment

29

Molecular Pharmaceutics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 30 of 40

Figure 2. 13C Solid-state NMR spectra: Panel A – Compound 1 (α polymorph - i, δ polymorph – ii, and amorphous - iii); Panel B – Compound 5 (Form I – i, Form III – ii, Form IV – iii, amorphous - iv); Panel C – Compound 2 (Form I – i, amorphous - ii); Panel D – Compound 3 (Form I – i, amorphous - ii); Panel E – Compound 4 (Form I – i, Form II – ii, amorphous - iii); Panel F (7 – i and 8 – ii).

ACS Paragon Plus Environment

30

Page 31 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

Table 2. The Chemical Shift Value (in ppm) of the Carboxylic Acid Carbon as Observed using 13 C Solid State NMR Spectroscopy. Compound

Solid state

Hydrogen-bonding synthon***

Carboxylic acid C=O chemical shift value (ppm)

1

α polymorph*

acid-acid

174.4

δ polymorph

acid-pyridine

169.1

Amorphous

acid-pyridine

169.9

Form I

acid-acid

174.0

Amorphous

acid-pyridine

169.6

Form I

acid-pyridine

169.0-170.4

Amorphous

acid-pyridine

169.6

Form I*

acid-acid

174.0

Form II

acid-acid**

174.3

Amorphous

acid-pyridine

170.1

Form I*

acid-pyridine

169.1

Form III

acid-acid

174.9

Form IV

acid-acid

172.7

Amorphous

acid-pyridine

169.5

7

Form I

acid-acid

177.7-176.3

8

Form I

acid-acid

174.7

9

Form I

acid-acid

N/A

2

3

4

5

*

Used as the starting material for DSC tests

**

Crystal structure not available. Identification of the synthon is based on SSNMR results.

***

In the amorphous state, it is the predominant hydrogen-bonding synthon.

ACS Paragon Plus Environment

31

Molecular Pharmaceutics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 32 of 40

Figure 3. Panel A: VT-FTIR spectra of compound 1: α polymorph (i), δ polymorph (ii), melt (iii), and amorphous phase (iv); Panel B: Compound 6 in crystalline state (i), and melt state (ii) and amorphous phase (iii); Panel C: Compound 7 in crystalline state (i), and melt state (ii); Panel D: Compound 8 in crystalline state (i), and melt state (ii).

ACS Paragon Plus Environment

32

Page 33 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

Table 3. Variable Temperature FTIR Data of the Model Compounds in Crystalline, Melt and Amorphous State Showing the Carbonyl and Aromatic Ring Vibrations. Compound

Solid-state

Carbonyl (C=O) vibration (cm-1)

1

α polymorph

1660.0 (sh)

2

δ polymorph melt amorphous Form I

1663.0 (sh) 1684.0 1679.0 1668.4

melt amorphous

1682.0 1678.0

3

Form I melt amorphous

1658.0 1684.0 1677.7

4

Form I

1675.0

melt

1681.0

amorphous

1680.0

Form I Form III

1677.0 1663.5

Form IV

1675.4

amorphous Form I melt amorphous

1679.1 1712.0 1700.0 1669.0

Form I melt Form I

1652.4 1656.1 1645.0

melt Form I melt

1647.0 1654.6 1656.0

5

6

7 8

9

Aromatic (C=C, C=N) ring vibrations (cm-1) 1610.2, 1585.0, 1570.0 1595.0, 1580.0 1588.0, 1572.0 1592.0, 1574.0 1615.3, 1600.2, 1584.4, 1517.7 1611.0, 1580.0 1613.0, 1583.0, 1553.0 1597.7, 1581.2 1580 .0 1611.1, 1582.7, 1554.8 1622.0, 1603.0, 1591.0, 1574.0 1616.0, 1600.0, 1586.0, 1574.0 1618.0, 1602.0, 1589.0, 1554.6 1593.5, 1578.9 1608.2, 1596.4, 1579.9, 1566.1 1615.8, 1594.4, 1580.0, 1566.1 1609.8, 1580.1 1585.0, 1565.0 1579.8, 1559.0 1623.0, 1581.0, 1560.0 1587.6, 1573.2 1585.3, 1570.4 1620.0, 1600.0, 1587.0 1618.0, 1600.0 1620.7, 1597.2 1584.0

Temperature used for data collection (°C) 25 25 165 25 25 175 25 25 180 25 25 165 25 25 25 25 25 25 140 25 25 195 25 240 25 220

sh = shoulder

ACS Paragon Plus Environment

33

Molecular Pharmaceutics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 34 of 40

Figure 4. Raman spectra of compound 1: α polymorph (a), δ polymorph (b), amorphous-state (c).

ACS Paragon Plus Environment

34

Page 35 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

Figure 5. Hirshfeld surfaces of acid-pyridine and acid-acid dimer of 1. The red spots indicate the contacts shorter than van der Waals distance, while blue indicates longer contacts.

ACS Paragon Plus Environment

35

Molecular Pharmaceutics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 36 of 40

Table 4. Hydrogen Bonding Energies of Compounds 1-3

Compound

∆E (–COOH···–COOH) (kJ/mol)

∆E (–COOH···N–) (kJ/mol)

1

−32.07

−49.64

2

−32.85

−53.15

3

−33.54

−50.70

ACS Paragon Plus Environment

36

Page 37 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

REFERENCES: 1.

2. 3. 4. 5. 6. 7. 8.

9. 10.

11.

12.

13.

14.

15.

16.

17.

Rodriguez-Spong, B.; Price, C. P.; Jayasankar, A.; Matzger, A. J.; Rodriguez-Hornedo, N. General principles of pharmaceutical solid polymorphism: a supramolecular perspective. Adv. Drug Deliv. Rev. 2004, 56, 241-74. Newman, A.; Knipp, G.; Zografi, G. Assessing the Performance of Amorphous Solid Dispersions. J. Pharm. Sci. 2012, 101, 1355-1377. Cui, Y. A material science perspective of pharmaceutical solids. Int. J. Pharm. 2007, 339, 3-18. Delori, A.; Friščić, T.; Jones, W. The role of mechanochemistry and supramolecular design in the development of pharmaceutical materials. CrystEngComm 2012, 14, 2350. Babu, N. J.; Nangia, A. Solubility Advantage of Amorphous Drugs and Pharmaceutical Cocrystals. Cryst. Growth Des. 2011, 11, 2662-2679. Hancock, B. C.; Parks, M. What is the true solubility advantage for amorphous pharmaceuticals? Pharm. Res. 2000, 17, 397-404. Yu, L. Amorphous pharmaceutical solids: preparation, characterization and stabilization. Adv. Drug Deliv. Rev. 2001, 48, 27-42. Kawakami, K.; Pikal, M. J. Calorimetric investigation of the structural relaxation of amorphous materials: Evaluating validity of the methodologies. J. Pharm. Sci. 2005, 94, 948-965. Bhugra, C.; Pikal, M. J. Role of thermodynamic, molecular, and kinetic factors in crystallization from the amorphous state. J. Pharm. Sci. 2008, 97, 1329-1349. Baird, J. A.; Van Eerdenbrugh, B.; Taylor, L. S. A Classification System to Assess the Crystallization Tendency of Organic Molecules from Undercooled Melts. J. Pharm. Sci. 2010, 99, 3787-3806. Descamps, M.; Dudognon, E. Crystallization from the Amorphous State: NucleationGrowth Decoupling, Polymorphism Interplay, and the Role of Interfaces. J. Pharm. Sci. 2014, 103, 2615-2628. Mahlin, D.; Ponnambalam, S.; Hockerfelt, M. H.; Bergstrom, C. A. S. Toward In Silico Prediction of Glass-Forming Ability from Molecular Structure Alone: A Screening Tool in Early Drug Development. Mol. Pharm. 2011, 8, 498-506. Alhalaweh, A.; Alzghoul, A.; Kaialy, W.; Mahlin, D.; Bergstrom, C. A. Computational predictions of glass-forming ability and crystallization tendency of drug molecules. Mol. Pharm. 2014, 11, 3123-32. Alhalaweh, A.; Alzghoul, A.; Mahlin, D.; Bergstrom, C. A. S. Physical stability of drugs after storage above and below the glass transition temperature: Relationship to glassforming ability. Int. J. Pharm. 2015, 495, 312-317. Wyttenbach, N.; Kirchmeyer, W.; Alsenz, J.; Kuentz, M. Theoretical Considerations of the Prigogine-Defay Ratio with Regard to the Glass-Forming Ability of Drugs from Undercooled Melts. Mol. Pharm. 2016, 13, 241-250. Kawakami, K.; Harada, T.; Miura, K.; Yoshihashi, Y.; Yonemochi, E.; Terada, K.; Moriyama, H. Relationship between Crystallization Tendencies during Cooling from Melt and Isothermal Storage: Toward a General Understanding of Physical Stability of Pharmaceutical Glasses. Mol. Pharm. 2014, 11, 1835-1843. Kawakami, K.; Usui, T.; Hattori, M. Understanding the glass-forming ability of active pharmaceutical ingredients for designing supersaturating dosage forms. J. Pharm. Sci. 2012, 101, 3239-3248.

ACS Paragon Plus Environment

37

Molecular Pharmaceutics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

18. 19. 20.

21.

22.

23. 24.

25.

26. 27. 28.

29.

30.

31.

32. 33. 34. 35.

Page 38 of 40

Turnbull, D. Under what conditions can a glass be formed? Contemporary Physics 1969, 10, 473-488. Uhlmann, D. R. A kinetic treatment of glass formation. J. Non-Cryst. Solids 1972, 7, 337348. Chelli, R.; Cardini, G.; Procacci, P.; Righini, R.; Califano, S. Molecular dynamics of glass-forming liquids: Structure and dynamics of liquid metatoluidine. J. Chem. Phys. 2002, 116, 6205-6215. Dirama, T. E.; Carri, G. A.; Sokolov, A. P. Role of hydrogen bonds in the fast dynamics of binary glasses of trehalose and glycerol: A molecular dynamics simulation study. J. Chem. Phys. 2005, 122, 8. Habgood, M.; Lancaster, R. W.; Gateshki, M.; Kenwright, A. M. The Amorphous Form of Salicylsalicylic Acid: Experimental Characterization and Computational Predictability. Cryst. Growth Des. 2013, 13, 1771-1779. Baran, J.; Davydova, N. A.; Pietraszko, A. Spectroscopic study of the formation of molecular glasses. Journal of Molecular Structure 2005, 744-747, 301-305. Lebel, O.; Maris, T.; Perron, M. E.; Demers, E.; Wuest, J. D. The dark side of crystal engineering: Creating glasses from small symmetric molecules that form multiple hydrogen bonds. J. Am. Chem.Soc. 2006, 128, 10372-10373. Plante, A.; Mauran, D.; Carvalho, S. P.; Page, J. Y. S. D.; Pellerin, C.; Lebel, O. Tg and Rheological Properties of Triazine-Based Molecular Glasses: Incriminating Evidence Against Hydrogen Bonds. J. Phys. Chem. B 2009, 113, 14884-14891. Wang, R. Y.; Pellerin, C.; Lebel, O. Role of hydrogen bonding in the formation of glasses by small molecules: a triazine case study. J. Mater. Chem. 2009, 19, 2747-2753. Meunier, A.; Lebel, O. A Glass Forming Module for Organic Molecules: Making Tetraphenylporphyrin Lose its Crystallinity. Org. Lett. 2010, 12, 1896-1899. Eren, R. N.; Plante, A.; Meunier, A.; Laventure, A.; Huang, Y. S.; Briard, J. G.; Creber, K. J.; Pellerin, C.; Soldera, A.; Lebel, O. One ring to rule them all: effect of aryl substitution on glass-forming ability in mexylaminotriazine molecular glasses. Tetrahedron 2012, 68, 10130-10144. Laventure, A.; Soldera, A.; Pellerin, C.; Lebel, O. Heads vs. tails: a double-sided study of the influence of substituents on the glass-forming ability and stability of aminotriazine molecular glasses. New J. Chem. 2013, 37, 3881-3889. Yu, L.; Reutzel-Edens, S. M.; Mitchell, C. A. Crystallization and polymorphism of conformationally flexible molecules: Problems, patterns, and strategies. Org. Process Res. Dev. 2000, 4, 396-402. Siniti, M.; Carre, J.; Letoffe, J. M.; Bastide, J. P.; Claudy, P. The thermal-behavior of hexitols .Part 1. vitrification and crystallization of iditol, mannitol, sorbitol and dulcitol. Thermochim. Acta 1993, 224, 97-104. Siniti, M.; Carre, J.; Bastide, J. P.; Letoffe, J. M.; Claudy, P. The thermal-behavior of hexitols . Part 2. polymorphism of dulcitol. Thermochim. Acta 1993, 224, 105-110. Siniti, M.; Jabrane, S.; Letoffe, J. M. Study of the respective binary phase diagrams of sorbitol with mannitol, maltitol and water. Thermochim. Acta 1999, 325, 171-180. Jabrane, S.; Letoffe, J. M.; Claudy, P. Vitrification and crystallization in the R (-)1,2propanediol-S (+)1,2-propanediol system. Thermochim. Acta 1995, 258, 33-47. Long, S. H.; Parkin, S.; Siegler, M. A.; Cammers, A.; L., L. T. Polymorphism and Phase Behaviors of 2-(Phenylamino)nicotinic Acid. Cryst. Growth Des. 2008, 8, 4006-4013.

ACS Paragon Plus Environment

38

Page 39 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

36.

37.

38.

39.

40. 41.

42.

43. 44.

45. 46. 47. 48.

Long, S. H.; Parkin, S.; Siegler, M.; Brock, C. P.; Cammers, A.; Li, T. L. Polymorphism of an organic system effected by the directionality of hydrogen-bonding chains. Cryst. Growth Des. 2008, 8, 3137-3140. Long, S. H.; Li, T. L. Controlled Formation of the Acid-Pyridine Heterosynthon over the Acid-Acid Homosynthon in 2-Anilinonicotinic Acids. Cryst. Growth Des. 2009, 9, 49934997. Long, S. H.; Li, T. L. Enforcing Molecule's pi-Conjugation and Consequent Formation of the Acid-Acid Homosynthon over the Acid-Pyridine Heterosynthon in 2-Anilinonicotinic Acids. Cryst. Growth Des. 2010, 10, 2465-2469. Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A., Jr.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, N. J.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, Ö.; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09, Revision C.01, Gaussian, Inc.: Wallingford CT, 2009. Dovesi, R.; Orlando, R.; Civalleri, B.; Roetti, C. S., V. R.; Zicovich-Wilson, C. M. Crystal09. Z. Kristallogr. 2005, 220, 571-573. Simon, S., Duran, M., Dannenberg, J. J. How does basis set superposition error change the potential surfaces for hydrogen bonded dimers? J. Chem. Phys. 1996, 105, 1102411031. Feng, S., Li, T. Predicting Lattice Energy of Organic Crystals by Density Functional Theory with Empirically Corrected Dispersion Energy. J. Chem. Theory Comput. 2006, 2, 149-156. Spackman, M. A.; Jayatilaka, D. Hirshfeld surface analysis. CrystEngComm 2009, 11, 19-32. Duddu, S. P.; Zhang, G.; Dal Monte, P. R. The relationship between protein aggregation and molecular mobility below the glass transition temperature of lyophilized formulations containing a monoclonal antibody. Pharm. Res. 1997, 14, 596-600. Kumar, S. S.; Nangia, A. A Solubility Comparison of Neutral and Zwitterionic Polymorphs. Cryst. Growth Des. 2014, 14, 1865-1881. Sbit, M.; Dupont, L.; Dideberg, O.; Liegeois, J. F.; Delarge, J. Structure of 2diphenylamino-carboxylic acid. Acta Crystallogr. C 1987, 43, 926-928. SeethaLekshmi, S.; Guru Row, T. N. Conformational Polymorphism in a Non-steroidal Anti-inflammatory Drug, Mefenamic Acid. Cryst. Growth Des. 2012, 12, 4283-4289. Andersen, K. V.; Larsen, S.; Alhede, B.; Gelting, N.; Buchardt, O. Characterization of 2 polymorphic forms of tolfenamic acid, n-(2-methyl-3-chlorophenyl)anthranilic acid their crystal-structures and relative stabilities. J. Chem. Soc., Perkin Trans. 2 1989, 14431447.

ACS Paragon Plus Environment

39

Molecular Pharmaceutics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

49.

50.

51.

52.

53. 54.

55. 56.

57. 58. 59. 60. 61. 62. 63. 64.

Page 40 of 40

Mattei, A.; Li, T. Polymorph formation and nucleation mechanism of tolfenamic acid in solution: an investigation of pre-nucleation solute association. Pharm. Res. 2012, 29, 460-70. Mahlin, D.; Ponnambalam, S.; Hockerfelt, M. H.; Bergstrom, C. A. Toward in silico prediction of glass-forming ability from molecular structure alone: a screening tool in early drug development. Mol. Pharm. 2011, 8, 498-506. Gradys, A.; Sajkiewicz, P.; Minakov, A. A.; Adamovsky, S.; Schick, C.; Hashimoto, T.; Saijo, K. Crystallization of polypropylene at various cooling rates. Mater. Sci. Eng., A 2005, 413, 442-446. Bouzidi, L.; Narine, S. S. Evidence of Critical Cooling Rates in the Nonisothermal Crystallization of Triacylglycerols: A Case for the Existence and Selection of Growth Modes of a Lipid Crystal Network. Langmuir 2010, 26, 4311-4319. Aliev, A. E.; Harris, K. D. M., Probing Hydrogen Bonding in Solids Using Solid State NMR Spectroscopy. 2004; Vol. 108. Jabeen, S.; Dines, T. J.; Leharne, S. A.; Chowdhry, B. Z. Raman and IR spectroscopic studies of fenamates--conformational differences in polymorphs of flufenamic acid, mefenamic acid and tolfenamic acid. Spectrochim. Acta. A Mol. Biomol. Spectrosc. 2012, 96, 972-85. Gotze, W.; Sjogren, L. Relaxation processes in supercooled liquids. Rep. Prog. Phys. 1992, 44, 241-376. Murdande, S. B.; Pikal, M. J.; Shanker, R. M.; Bogner, R. H. Aqueous solubility of crystalline and amorphous drugs: Challenges in measurement. Pharm. Dev. Technol. 2011, 16, 187-200. Wuest, J. D.; Lebel, O. Anarchy in the solid state: structural dependence on glass-forming ability in triazine-based molecular glasses. Tetrahedron 2009, 65, 7393-7402. Wang, R.; Pellerin, C.; Lebel, O. Role of hydrogen bonding in the formation of glasses by small molecules: a triazine case study. J. Mater. Chem. 2009, 19, 2747. Shintani, H.; Tanaka, H. Frustration on the way to crystallization in glass. Nature Physics 2006, 2, 200-206. Tanaka, H. Two-order-parameter description of liquids. II. Criteria for vitrification and predictions of our model. J. Chem. Phys. 1999, 111, 3175-3182. Tanaka, H. Two-order-parameter model of the liquid–glass transition. II. Structural relaxation and dynamic heterogeneity. J. Non-Cryst. Solids 2005, 351, 3385–3395. Davey, R. J.; Schroeder, S. L.; Ter Horst, J. H. Nucleation of organic crystals - a molecular perspective. Angew. Chem. Int. Ed. Engl. 2013, 52, 2166-79. Kashchiev, D.; Van Rosmalen, G. M. Review: Nucleation in solutions revisited. Cryst. Res. Technol. 2003, 38, 555-574. Tanaka, H. Relationship among glass-forming ability, fragility and short-range bond ordering of liquids. J Non-Cryst Solids 2005, 351, 678-690.

ACS Paragon Plus Environment

40