The Crystal Engineering of Two Anti- bacterial Pharmaceutical Ingredients: Cocrystals and Molecular Salts of a Series of Sulfa-Drugs and the Polymorphism and Cocrystals of an Isoniazid Derivative. By Matthew Clarke Scheepers Submitted in accordance with the requirements for the degree of Doctor of Philosophy In the subject Chemistry At the UNIVERSITY OF THE WITWATERSRAND Supervisor: Prof. A. Lemmerer January 2023 i DECLARATION I declare that “The Crystal Engineering of Two Anti-bacterial Pharmaceutical Ingredients: Cocrystals and Molecular Salts of a Series of Sulfa-Drugs and the Polymorphism and Cocrystals of an Isoniazid Derivative.” is my own unaided work and that all sources I have used or quoted have been indicated and acknowledged by means of complete references. It has not been submitted before for any degree or examination at any other University. 31 January 2023 Matthew Clarke Scheepers Date ii Abstract Crystal engineering is the design and synthesis of new solid forms by using the knowledge of intermolecular interactions and crystal packing. The goal of crystal engineering is to improve the properties of materials without altering the chemical identity of the materials. This work can be divided into three major parts. The first part deals with a series of sulfa drugs and its cocrystals. The second part deals with exploring the cocrystals of 3,5-dinitrobenzoic acid, in particular using Hirshfeld surfaces. The last part deals with the polymorphism of an isoniazid derivative and the cocrystals formed with it. In all cases these new solids were characterized by single crystal X-ray diffraction (SC-XRD), powder X-ray diffraction (PXRD), differential scanning calorimetry (DSC), thermal gravimetric analysis (TGA), and Fourier-transform infrared spectroscopy (FTIR). When designing new solid forms (cocrystals and molecular salts) it would be ideal to use a knowledge based system on the use of synthons. Sometimes this knowledge is limited and can hinder future work. Sulfa drugs such as sulfamethazine (sz), sulfapyridine (sp), sulfathiazole (st) and sulfamethoxazole (sm) have been shown to exhibit unique solid-state behaviour such as cocrystal-salt polymorphism and tautomerism. The number of known cocrystals and molecular salts consisting of the mentioned drugs are limited, which hinders research into exploring this interesting solid-state behaviour. Therefore, the structural landscape of these sulfa drugs was expanded in hopes of observing new cases where this solid-state behaviour is observed. Cocrystals and molecular salts of these sulfa drugs were synthesized using various techniques, such as solvent evaporation and mechanochemical grinding (both dry and liquid- assisted grinding). Twenty-five different coformers were chosen to use for these cocrystallization experiments, most of which are either benzoic acid/benzoic acid derivatives or a pyridine derivative. sz formed eight successful cocrystals and one molecular salt with benzoic acid and its derivatives. SC-XRD showed that eight of the coformers that interacted iii with sz formed the sulfonamide-carboxyl synthon; the only exception to this was sz + 4- hydroxybenzoic acid, which the sulfonimide-carboxyl synthon formed instead. Five cocrystals and one molecular salt containing st, three cocrystals and one molecular salt of sm, and four cocrystals and one molecular salt containing sp were also synthesized. Most of these included using 2-aminopyridine or one of its derivatives. The second part deals with the cocrystals and molecular salts of 3,5-dinitrobenzoic acid (dnba). dnba is often used as a popular coformer for crystal engineering purpose. However, very little work centered on dnba with coformers has been reported. In this work we report new multi- component crystals containing dnba, which include one hydrate, one solvate, one molecular salt, and four cocrystals. The coformers include: 2-acetylpyridine, 3-cyanopyridine, flufenamic acid, 4-dimethylaminobenzophenone, pyridoxine, theophylline, and thiourea. In addition to the strong hydrogen bonding expected, several weaker intermolecular interactions were identified using Hirshfeld surfaces, which included C−H···π bonding, π-hole, and π···π interactions. The Hirshfeld surfaces indicated that these weaker interactions had a significant effect on the packing of these multi-component crystals. These multi-component crystals were compared with the crystal structures reported in the Cambridge structural database (CSD), which has given some significant insights in the structural landscape of dnba. Isoniazid (inh) is a simple, useful active pharmaceutical ingredient (API) used to treat Myobacterium tuberculosis. It is composed of a hydrazine and pyridine ring. The hydrazine group can be modified in a multitude of ways, including using a Schiff-base condensation reaction. This is achieved by using an appropriate ketone or aldehyde. In this work isoniazid was derivitised using diacetone alcohol (4-hydroxy-4-methyl-2-pentanone), generating N’- [(2E)-4-hydroxy-4-methylpentan-2-ylidene]pyridine-4-carbohydrazide (iz4h4m2p). This derivative was found to be polymorphic, existing in two forms, form I and form II. The iv difference between these two forms is distinguished by the major hydrogen bond pattern, which is either a chain hydrogen bond motif formed between the hydroxyl group and amide groups (the pyridine ring is not involved) or a ring based hydrogen bond motif that forms dimers with the hydroxyl group forming a hydrogen bond to the pyridine ring. Form I was found to be metastable with respect to form II, with form I converting to form II upon heating before melting. Form II does not convert to form I. In addition to exploring the polymorphism of iz4h4m2p, several cocrystals of iz4h4m2p were synthesized. The coformers chosen were benzoic acid derivatives. Most of these cocrystals formed a hydrogen bond between the carboxylic acid and pyridine ring. This left the hydroxyl group of iz4h4m2p to form a hydrogen bond to the amide group, forming a chain hydrogen bond motif similar to the one observed in form I. The only deviation from this was observed in cocrystals where the benzoic acid included a hydroxyl group of its own, such as 2,5-dihydroxybenzoic acid. v Acknowledgements First, I would like to thank my supervisor, Prof. Andreas Lemmerer, for his unwavering support, dedication, patience and kindness. Under your supervision, not only have I have grown my knowledge and skills but I have also grown as a person. I also thank you for allowing me to attend a few conferences and the crystallographic school that gave me insight into the nature of research in crystallography. I would also like to thank Prof. Manuel Fernandes and Prof. Demetrius Levendis for their time, advice and kindness. I would also like to thank the rest of the structural chemistry group over the past 6 years for their support, including Cara, Prathapa, Delbert, Attiyah and Tracy. Lastly I would like to thank my family for their endless support and unconditional love, making it possible for me to achieve my hopes and dreams. Poster Poster presented: Synthesis and characterization of a series of multicomponent crystals with sulfanilamide derivatives, presented at the 11th Bologna’s convention on crystal forms held online 10-11 September 2021 vi The structure and outputs of the Thesis There are several styles present in this work. The introductory, experimental and concluding chapters have been written in a uniform style. The central chapters (Chapters 3-5) have been published in different journals, each with its own respective style required for that journal. The published work has been presented as the essential research chapters since the significant findings from a large amount of work have been presented in a concise way. In chapters 3-5 a brief description of the papers is present preceding the papers. vii Contributions from the authors There were five papers published for this thesis. The conceptual ideas, as well as most of the experimental work and analysis was developed and carried out by myself. I wrote the papers largely by myself. Prof. Andreas Lemmerer provided supervision and academic insight. Prof. Manuel Fernandes carried out the PIXEL calculations presented in chapter 5, which was requested by the reviewers of the paper, which he helped write the related parts in that paper. This was his only contribution. 1 Contents Chapter 1: The Crystal Engineering of Active Pharmaceutical Ingredients as an attempt to improve solid-state properties.............................................................................................................. 6 1.1 Crystals ........................................................................................................................................ 6 1.1.1 Introduction .......................................................................................................................... 6 1.1.2 Crystal Growth ..................................................................................................................... 8 1.2. Crystal Engineering and Supramolecular Chemistry .......................................................... 10 1.2.1 Introduction ........................................................................................................................ 10 1.2.2 Intermolecular interactions ............................................................................................... 10 1.2.3 Hydrogen Bonding ............................................................................................................. 11 1.2.4 π···π stacking ...................................................................................................................... 12 1.2.5 Polymorphism .................................................................................................................... 13 1.2.6 Cocrystals and Molecular Salts ........................................................................................ 19 1.2.7 Tautomerism ...................................................................................................................... 24 1.3 Hirshfeld Surfaces ..................................................................................................................... 25 1.4 Sulfa Drugs ................................................................................................................................ 30 1.4.1 Sulfapyridine ...................................................................................................................... 32 1.4.2. Sulfathiazole ...................................................................................................................... 33 1.4.3. Sulfamethoxazole .............................................................................................................. 34 1.4.4. Sulfamethazine .................................................................................................................. 35 1.5 Isoniazid ..................................................................................................................................... 35 1.5.1 Introduction to isoniazid ................................................................................................... 35 1.5.2 Blocking the hydrazine group ........................................................................................... 36 1.5.3 Polymorphism of isoniazid and its derivatives/cocrystals .............................................. 37 1.6 Aims and Objectives ................................................................................................................. 37 Chapter 2 Experimental Procedures and Aims and Objectives ..................................................... 39 2.1 Cocrystal/molecular salt synthetic methods ........................................................................... 39 2.2 Synthesis of iz4h4m2p ............................................................................................................... 40 2.3 Single Crystal X-Ray Diffraction............................................................................................. 40 2.4 Powder X-Ray Diffraction........................................................................................................ 41 2.5 Infrared Spectroscopy .............................................................................................................. 41 2.6 Differential Scanning Calorimetry .......................................................................................... 41 2.7 Hirshfeld Surface Analysis ....................................................................................................... 42 2.8 Hot Stage Microscopy ............................................................................................................... 42 2.9 Thermalgravimetric Analysis .................................................................................................. 42 2 Chapter 3 Synthesis and Characterization of a series of Sulfamethazine multi-component crystals with Various Benzoic Acids.................................................................................................. 43 3.1 introduction ............................................................................................................................... 43 Chapter 4 In Pursuit of Multicomponent Crystals of the Sulfa Drugs Sulfapyridine, Sulftahiazole and Sulfamethoxazole .................................................................................................. 55 4.1 Introduction ............................................................................................................................... 55 Chapter 5 Exploring the Crystal Structure Landscape of 3,5-Dinitrobenzoic Acid through Various Multicomponent Molecular Complexes.............................................................................. 81 5.1 Introduction ............................................................................................................................... 81 Chapter 6 Chains or rings? Polymorphism of an isoniazid derivative derivatized with diacetone alcohol .................................................................................................................................................. 95 6.1 Introduction ............................................................................................................................... 95 Chapter 7 Design of a series of cocrystals featuring isoniazid modified with diacetone alcohol103 7.1 Introduction ............................................................................................................................. 103 Chapter 8 Conclusions ...................................................................................................................... 115 References .......................................................................................................................................... 118 Appendix ............................................................................................................................................ 125 3 List of Figures Figure 1.1.1 The 14 Bravais lattices. Figure 1.1.2 Schematic illustration of the size dependence of the energetics of nucleation. Figure 1.2.4.1. The three different types of π···π stacking: (a) cofacial parallel packing, (b) parallel displaced packing and (c) Edge-to-face T-shaped packing. Figure 1.2.4.2 Graphical representation explaining why a cofacial parallel stack is unfavourable to parallel displaced, and a case where cofacial parallel stack might be feasible. Figure 1.2.5.1 Enantiotropic vs. monotropic energy-temperature diagrams Fig. 1.2.5.2 Schematic of the reaction coordinates for the crystallization of a dimorphic system. Scheme 1.2.7. The tautomerism of adenine: (A – major tautomer of adenine) hydrogen bonded to thymine (T) on the left versus adenine’s minor tautomer (A’) hydrogen bonded to cytosine (C) on the right. Figure 1.3.1 A plain surface showing the definition of di and de of 4-hydroxybenzoic acid (Refcode JOZZHI01). Figure 1.3.2 The Hirshfeld surface of 4hba (Refcode JOZZHI01) showing (a) di, (b) de, (c) dnorm, (d) the shape index and (e) the curvedness. Figure 1.3.3 A fingerprint plot of 4hba (Refcode JOZZHI01). Scheme 1.4. The basic structure of sulfonanilamide. Scheme 1.4.1 Basic structure of sulfapyridine showing its tautomerism. Scheme 1.4.2 The structure of st showing the sulfonamide tautomer. Scheme 1.4.3 Structure of sm. Scheme 1.4.4. Structure of sz showing the sulfonamide tautomer (right) and sulfonimide tautomer (left). Scheme. 1.5.3. General scheme of the Schiff-base condensation reaction of isoniazid with a ketone or aldehyde (R1 = H). Scheme 2.2 Reaction scheme for the synthesis of iz4h4m2p. List of Tables Table 2.1 List of coformers used in this work. 4 5 List of Abbreviations 2ap: 2-aminopyridine 3ap: 3-aminopyridine 4ap: 4-aminopyridine 2a3np: 2-amino-3-nitropyridine 2a5np: 2-amino-5-nitropyridine 2a5clp: 2-amino-5-chloropyridine bipy: 4,4’-bipyridine BCS: Biopharmaceutics Classification System 2hba: 2-hydroxybenzoic acid (salicylic acid) 3hba: 3-hydroxybenzoic acid 4hba: 4-hydroxybenzoic acid 2hp: 2-hydroxypyridine 3hp: 3-hydroxypyridine 4hp: 4-hydroxypyridine 2c4n: 2-chloro-4-nitrobenzoic acid 2c5n: 2-chloro-5-nitrobenzoic acid 3cnp: 3-cyanopyridine 4cnp: 4-cyanopyridine ca: cinnamic acid CSD: Cambridge structural database dnba: 3,5-dinitrobenzoic acid dcba: 3,4-dichlorobenzoic acid DSC: Differential Scanning Calorimetry FTIR: Fourier-transforminfrared spectroscopy 4hba: 4-hydroxybenzoic acid inh : isoniazid iz4h4m2p: N'-[(2E)-4-hydroxy-4-methylpentan-2-ylidene]pyridine-4-carbohydrazide LAG: Liquid-assisted grinding PXRD: Powder X-Ray Diffraction PABA: 4-aminobenzoic acid sa: succinic acid sac: saccharine SC-XRD: Single Crystal X-Ray Diffraction sm : sulfamethoxazole sp : sulfapyridine st : sulfathiazole sz : sulfamethazine ta: Toluic acid 6 Chapter 1: The Crystal Engineering of Active Pharmaceutical Ingredients as an attempt to improve solid-state properties 1.1 Crystals 1.1.1 Introduction A crystal is a solid made up of a periodic and ordered arrangement of atoms, ions and/or molecules in three-dimensional space known as a lattice. The crystal structure is best represented by using a unit cell, which can replicate the entire crystal structure by using only translation. Depending on the lattice, several possible unit cells may be defined, however, only one unit cell is considered. The one chosen is one that best represents the different symmetries that are present in the crystal structure while minimizing the size required to recreate the entire crystal structure. All crystals exist in one of seven possible crystal systems. These crystal systems are: triclinic, monoclinic, orthorhombic, tetragonal, trigonal, hexagonal and cubic. These seven crystal systems are then classified further as being part of one of fourteen Bravais lattices (Fig. 1.1.1). These are further classified by space groups, which presents all the symmetry present in the crystal structure. Symmetry operations include inversion, rotations, mirror planes, screw axes and glide planes, in addition to the identity operation E. There are 230 space groups in total, with each space group named after the major symmetry operation(s) present in each crystal structures. 7 Figure 1.1.1 The 14 Bravais lattices (images adapted from Wikipedia page1). For molecular crystals consisting of organic compounds, approximately 75% of reported organic crystals appears in just five space groups, which are: P21/c, 𝑃1̅, P21, P212121 and C2/c.2 8 Many solids of interest form in a crystal, and the study of crystals forms the basis of many areas of research. The properties of solids arise from the arrangement of atoms of molecules in the solid state, as well as the interactions between atoms and/or molecules present in the solid state. Therefore, in order to understand how a material behaves, the solid state structure needs to be first determined, and understood from an atomistic or molecular viewpoint. 1.1.2 Crystal Growth A standard method for purifying organic compounds typically involves using recrystallization. Here, the impure compound is dissolved in a suitable solvent. This solution is then filtered (leaving behind insoluble impurities) and then evaporated (either through heating or slow evaporation). The process then ultimately leaves behind crystals of the (hopefully) pure compound. A series of questions remains: “How did these crystals form?” “How did the individual molecules come together to form these crystals?” Crystallization is the process where atoms, molecules or ions aggregate together to form crystals. Crystallization is governed by two steps: nucleation and growth. Nucleation is the first step where a first order phase transition occurs, that is the change from being part of an aqueous solution, liquid or gas phase transforms into the solid phase. Two different types of nucleation exists: heterogeneous nucleation and homogenous nucleation. Homogenous nucleation occurs in the absence of the influence of foreign materials, for example, nucleation that occurs in the bulk or “centre” of a solution (not necessarily aqueous). Heterogeneous nucleation occurs at the surface of another material, for example the walls of glassware. Heterogeneous nucleation is the most dominant form of nucleation, which serves as the rationale for scratching the surface of glass walls to promote crystal growth. In order for nucleation to occur, the Gibbs free energy for the pre-existing phase (solution, liquid, gas) must become greater than the Gibbs free energy of the emerging phase.3,4 The Gibbs energy of the emerging nucleus is dependent on two 9 factors. The first is the Gibbs free energy per unit volume, i.e. the bulk. This energy is always negative. The second is related to the surface energy of the emerging nucleus, which is always positive. Fig. 1.1.2 represents these two energies. Small nuclei will consist of being more of a surface than consisting of bulk. These molecules will be loosely bound to the rest of the nucleus, which will result in the destabilisation of the nucleus and thus its dissolution. As more molecules aggregate to the developing nucleus, its bulk increases. At the critical radius rcrit, the Gibbs free energy between the interfacial and bulk energy reaches a maximum. Additional molecules at this point decreases this Gibbs free energy, which results in the growth of the nucleus. Heterogeneous nucleation is more prevalent than homogenous nucleation because the surface of the growing nucleus is substantially smaller than that of a free-growing nucleus, which ultimately causes crystal growth to occur using smaller nuclei. Figure 1.1.2 Schematic illustration of the size dependence of the energetics of nucleation. 10 1.2. Crystal Engineering and Supramolecular Chemistry 1.2.1 Introduction Organic chemists often use the process of recrystallization to purify organic samples. This involves the assembly of a random collection of molecules in solution to form an organized assembly that results in a crystal. If chemistry is about the control of chemical processes, is it possible to control the crystallization of molecules? Is it possible to change the assembly of molecules? Is there any benefit in doing so? The answer to these questions is yes, it is possible to manipulate the crystallization process (to an extent, at least), which forms part of the field of crystal engineering. Crystal engineering can be defined as the “understanding of intermolecular interactions in the context of crystal packing and in the utilization of such understanding in the design of new solids with desired physical and chemical properties.”5 This is achieved through the understanding of the intermolecular interactions between atoms, molecules and ions, which ultimately affects how they will pack in the solid state. 1.2.2 Intermolecular interactions The arrangement of molecules in crystal structures is driven by molecular recognition. Molecular recognition is the aspect where certain intermolecular interactions are favoured and co-ordination of molecules occurs due to these interactions. Non-covalent interactions such as hydrogen bonding, π∙∙∙π, metal co-ordination (chelation), van der Waals and electrostatic interactions often dictate the overall geometry and direction molecules form in the network. For e.g., DNA consists of two nucleobase strands held together by hydrogen bonds.6 It was the recognition that the nucleobases can only bond with its counterpart (eg. adenine with thymine, guanine with cytosine) that eventually led to elucidation of the overall structure determination 11 of DNA and hence its function.6 The understanding of the possible interactions can allow one to gain insight into how molecules form their respective crystal structures and hence predict/explain the resulting physical properties. 1.2.3 Hydrogen Bonding The hydrogen bond is one of the most important intermolecular interactions to consider in the design of molecular crystals. The IUPAC definition for the hydrogen bond is as follows: The hydrogen bond is an attractive interaction between a hydrogen atom from a molecule or a molecular fragment X–H in which X is more electronegative than H, and an atom or a group of atoms in the same or a different molecule, in which there is evidence of bond formation. 7 If the hydrogen bond is represented as X—H···A, X—H is defined as the hydrogen bond donor and A as the hydrogen bond acceptor. It is possible for a hydrogen bond donor or hydrogen bond acceptor to form more than one hydrogen bond simultaneously. If an atom forms two hydrogen bonds (either as a hydrogen bond donor or a hydrogen bond acceptor), the arrangement is referred to as a bifurcated hydrogen bond. Cases where an atom shares three hydrogen bonds is referred to as a trifurcated hydrogen bond, although such cases are rare.8 The bond energies of hydrogen bonds are typically in the range of 10-65 kJ mol-1, which are significantly higher in energy than that of the van der Waals forces (<8 kJ mol-1).9 The length of the hydrogen bond (X···A) can vary between 2.45 – 3.35 Å, while hydrogen bond angles (∠ X—H···A) can range between 90° - 180°. 8,10 The hydrogen bond is a directional force in molecular crystals, which is significant enough that a formal notation based on Graph Sets can be used to describe the arrangement of the molecule based on the hydrogen bonds present. These patterns are defined according to the number of hydrogen bond donors, acceptors and the number of atoms involved in making up the motif. 12 The patterns in which hydrogen bonding forms between neighbouring molecules may be described as one of four pattern types: chains (C), Rings (R), Discrete (D) or Self (S).11 Graph sets are presented in the following form 𝑮𝒅 𝒂(𝒏), where G represents the pattern, a represents the total number of hydrogen bond acceptors, d the total number of hydrogen bond donors and n represents the total number of atoms involved. The reasoning for using such definitions is that hydrogen bonding can be reduced from a complex pattern to that of a simple descriptive one, such that the resulting properties from such bonding can be explained with ease, and to attempt to predict such complex patterns. 1.2.4 π···π stacking Many organic molecules feature aromatic systems. These rings consists of a series of overlapping π-orbitals, which cause the electrons to be delocalized. It has been observed that there is an electrostatic interaction that exists between two or more rings, although the exact nature of this interaction is still not fully understood.12 π···π stacking forms part of several different areas of research, including biology and physics.12,13 Three main types of π···π stacking exists: edge-to-face T-shaped, parallel displaced, and cofacial parallel stacked (Fig. 1.2.4.1). π···π stacking is the result of a combination of both dispersion and electrostatic effects. Hunter and Sanders proposed a simple model for π···π stacking where a positively charged σ- framework is sandwiched between two negatively charged π-systems.13 Edge-to-face T shaped and parallel displaced π···π stacking systems are the most common, since most aromatic systems contain a partially positive hydrogen covalently bonded to the aromatic carbon ring. These partially positive hydrogens can form an attractive interaction with the partially negative π-systems. Cofacial parallel stacked remains the rarest of the three since the negatively charged π-systems will most likely repel each other. The only case where a cofacial parallel π···π stacking system could exist is where one of the coformers has strong electron withdrawing 13 functional groups present, such as the case of the cocrystal containing hexafluorobenzene and benzene (Fig. 1.2.4.2).14 Figure 1.2.4.1. The three different types of π···π stacking: (a) cofacial parallel packing, (b) parallel displaced packing and (c) Edge-to-face T-shaped packing. Figure 1.2.4.2 Graphical representation explaining why a cofacial parallel stack is unfavourable to parallel displaced, and a case where cofacial parallel stack might be feasible. (Graphics adapted from Chemistry Libre texts)15 1.2.5 Polymorphism 1.2.5.1. Introduction Polymorphism refers to the phenomenon where molecules or ions can exist in more than one crystal form, either through different conformations of the molecules (conformational polymorphism) or packing arrangements (packing polymorphism).16 Allotropy is where elements can have different solid state forms (such as diamond and graphite for carbon). Several thousand cases of polymorphism have been reported in the literature, with crystal 14 structures of organic and organometallic systems stored in the Cambridge Structural Database (CSD).17 Cocrystals can also exhibit polymorphism.18 The result of polymorphism is that different properties can arise. This can range from appearance (colour, shape etc.) to different electronic and thermal properties.19 In the pharmaceutical industry, polymorphism plays a major role in the research and development of API’s. Properties such as the dissolution rate and the solubility affect the bioavailability and therefore the effectiveness of the API.20 Since each polymorph will have properties that differ between each other, it becomes critical to determine and characterize all possible forms. Polymorphs also offer the advantage of becoming new intellectual property. McCrone once claimed that all compounds are polymorphic, and that the number of polymorphs discovered is proportional to the time and effort placed in finding them.21 Over the course of the last few decades it became apparent that polymorphism is the norm instead of being the exception. However, the number of polymorphs reported for different compounds can differ drastically. Some compounds such as salicylic acid and sucrose have only one form reported at ambient conditions, while other compounds such as 5-methyl-2-[(2-nitrophenyl)- amino]thiophene-3-carbonitrile, also known as ROY (Red, Orange, Yellow, named after the colours its polymorphic forms can exhbit), can have more than nine forms reported.22 Some compounds can be identified as monomorphic for several years before its first polymorphic form can be discovered, such as isoniazid23 or aspirin.24,25 It is even possible for different polymorphs to arise from the same crystallization event, a phenomenon known as concomitant polymorphism.21,26,27 There are cases where a polymorphic form that has been reported previously cannot be obtained from using the same conditions as before, a phenomenon known as “disappearing polymorphs.” The impact of polymorphism has only become more relevant over the past few decades.21,27 For example, the protease inhibitor Ritonavir, a drug used to treat HIV was initially marketed by Abbott (now AbbVie) in a capsule form that was sold as 15 Norvir oral liquid and Norvir semi-solid capsules.28 This first form (Form I) could easily dissolve in the desired water/ethanol mix. However, in 1998 a new, more thermodynamically stable form (Form II) appeared, which was identified from batches that failed the dissolution tests. As a result the new batches threatened the supply of this drug on the market as the new form could not be processed into the desired product, which resulted in massive recalls and became a major loss for the company. 1.2.5.2 Kinetics and Thermodynamics of polymorphism How does a crystal crystallise in a particular form? What are the conditions for one form to be favoured over the other? The formation of a particular form over the other will depend on both thermodynamic and kinetic factors. These factors play a role in both the initial crystallization process (from nucleation) and in solid-to-solid phase transformations. When comparing polymorphic forms, it is typical that one form is more thermodynamically stable than the others at some given temperature. The most thermodynamically stable polymorphic form is called the thermodynamic form while polymorphic forms that are more unstable than the thermodynamic form are referred to as metastable forms. It is expected that any metastable form will convert to the more thermodynamic form. The rate at which the metastable form transforms to the thermodynamically stable form is determined by kinetic factors. The rate at which this occurs can be fast (to the point as if the metastable form never existed) to slow (to the point where it seems as if nothing is happening). For example, diamond and coal are two allotropes of carbon, with diamond being the metastable phase and graphite the thermodynamically favoured form. It is expected that diamond should transform to graphite but the rate at which it does is so slow that practically the transformation does not occur. It is possible that under heating one polymorphic form will convert to another at some temperature. Consider a compound that consists of exactly two polymorphic forms: Form I and 16 Form II. The different polymorphic forms of molecular crystals are named using either a numeric approach (1,2,3… or I, II, III…) or using the Greek alphabet (α, β, γ…).29 Typically forms are named in order of discovery, but otherwise they are ordered in terms of melting points starting with the form with the lowest melting point. Let’s say that for our hypothetical compound we initially obtained Form I, which converts to Form II at some transition temperature Tp, i.e. a typical dimorphic system. If the transition of Form I to Form II is reversible at Tp, the process is enantiotropic. If the process is irreversible, it is monotropic. It should be noted that some phase transitions can be enantiotropic despite being irreversible due to kinetic reasons. Therefore, in order to decide conclusively if the phase transition is enantiotropic or monotropic, several rules need to be considered. Figure 1.2.5.1 Enantiotropic vs. monotropic energy-temperature diagrams The first rule to consider is the Heat-of-transition rule. It is stated as follows: “If an endothermal transition is observed at some temperature it may be assumed that there is a transition point below it, i.e. the two forms are related enantiotropically.” 17 “If an exothermal transition is observed at some temperature it may be assumed that there is no transition point below it, i.e. the two forms are either related monotropically or the transition temperature is higher.”30 This rule was derived based on the observation that ΔH and ΔS are usually positive, and that H curves do not intersect while the G curves intersects only once (Fig. 1.2.5.1). This rule has been obeyed by 99% of all polymorphic cases.30 In the case that the transition point cannot be measured easily, the Heat-of-fusion rule can be used. It states: “If the higher melting form has the lower heat of fusion the two forms are usually enantiotropic, otherwise they are monotropic.”30 This rule is obeyed if H curves follow the behaviour observed in Fig. 1.2.5.1. Deviation from this rule exists when the H curves diverge significantly or the difference between the melting points of both forms is more than 30 K.31 The success of this rule is essentially equivalent to the Heat-of-transition rule. In the case the Heat-of-fusion rule is unable to be used, the Entropy-of-fusion rule can be applied. The entropy of fusion is given by: Δ𝑆 = Δ𝐻𝑓 𝑇 Then the rule states that if the form with the higher melting point has the lower entropy of fusion then the two forms are enantiotropically related, otherwise they are monotropically related. 18 Kinetics plays a role in selecting the polymorphic form that crystallises first. Consider our dimorphic system in which either form can crystallise. There are two possible reaction pathways: one pathway has a low activation energy barrier but the overall free energy is higher while the other pathway has a higher activation energy barrier but the free energy is much lower (Fig. 1.2.5.2). Most often the metastable phase forms over the thermodynamic phase simply because the activation energy is lower. During nucleation, the phase boundary (between the surface of the nucleus and the surrounding environment) is associated with an increase in free energy. The system in turn, must overall decrease in free energy, which is achieved by increasing the size of the nucleus. The observation in which the metastable phase forms first due to having a lower activation barrier was first made by Ostwald, which he stated in his “Rule of Stages.” Although it is apparently true for many cases, exceptions arise such as in the case of concomitant polymorphism. It also becomes problematic when describing systems which are “apparently monomorphic.” 19 Fig. 1.2.5.2 Schematic of the reaction coordinates for the crystallization of a dimorphic system. 1.2.6 Cocrystals and Molecular Salts Multi-component complexes such as co-crystals and molecular salts are systems designed to contain more than one component in the solid state, and are a rapidly growing area of interest in modern chemistry. Cocrystals are the combination of two or more neutral molecules that interact with each other in a way that makes up the supramolecular complex.32 This distinguishes them from molecular salts where a proton transfer has occurred, leading to the formation of an ionic bond. In addition, cocrystals are further distinguished from hydrates and solvates in the sense that all the co-formers involved must be solid at ambient conditions. It should also be noted the spelling “cocrystal” and “co-crystal” are used by different authors and 20 journals and refer to the same thing, although agreement on which term to use has not been made. 1.2.5.1 pKa and the ΔpKa Rule As stated previously, molecular salts are solids where proton transfer as occurred. This results in the formation of a set of ions (cation and anion), with both species held together by an electrostatic attraction. It becomes important to consider the acidic and basic nature of many compounds in the crystal engineering of organic molecules. Many organic compounds are acids and bases, although most are weak in nature. In solution, both ionized and non-ionized species exist in equilibrium. For some generic acid HA, this equilibrium is expressed as: 𝐻𝐴 ⇌ 𝐻+ + 𝐴− For which the acid dissociation constant Ka can be represented as: 𝐾𝑎 = [𝐻+][𝐴−] [𝐻𝐴] One useful means of comparing the extent to which an acid may be deprotonated (or a base protonated) is through the use of the pKa which is defined as follows: 𝑝𝐾𝑎 = −log 𝐾𝑎 The pKa can also be used for bases by using the conjugate acid. Strong acids will have small or negative pKa’s while strong bases will have large positive pKa’s. Amphoterism is the case where a compound can act both as an acid or a base, and as such these compounds can have multiple pKa’s. An acid or base in solution will change the pH of the solution. This change is given by the Henderson-Hasselbalch equation below: 𝑝𝐻 = 𝑝𝐾𝑎 + log [𝐴−] [𝐻𝐴] 21 The Henderson-Hasselbalch equation indicates that in solution a mixture of an acid and its conjugate base (or a base and its conjugate acid) will exist, and the pH will affect the amount of acid (or conjugate acid) relative to its conjugative base (or base). The pKa is often used to compare the strength of an acid or a base with other acids and bases. The pKa of a compound is usually listed in tables or databases alongside many others. Although useful, there are several limitations to this. The first is that the pKa of many compounds are reported only for using water as the solvent. This presents a problem since different solvents have different effects on the solute. For example, according to a paper by Rossini et al.,33 which examines both experimental and theoretical pKa’s of various compounds in different solvents, the pKa of benzoic acid in water, acetonitrile, dimethyl sulfoxide and methanol is respectively 4.22, 20.10, 11.10 and 9.40. This makes sense, as solvents like water and methanol are protic solvents which will facilitate the proton transfer easier than an aprotic solvent like acetonitrile. This becomes problematic when the pKa for a compound in a particular solvent has not been determined experimentally (or even theoretically). Another issue is temperature. The pKa of many compounds has been reported at standard temperature (25 °C). Although the pKa does not change drastically within a few degrees, it becomes more significant at much higher temperatures. For example, solutions are often heated to allow complete dissolution of the solute, in which case the pKa will change. This change is determined according to the van't Hoff equation below: 𝑑 𝑑𝑇 ln 𝐾 = ∆𝐻𝑜 𝑅𝑇2 Where K is the equilibrium constant, R is the universal gas constant and ΔHo is the standard enthalpy. Accordingly, heating the solution will increase the Ka, which decreases the pKa. These factors need to be considered for solution-based cocrystallization strategies. 22 How does one predict whether a salt or a co-crystal will form (and if they form)? In order for a salt to form, in the case of most organic systems, a proton needs to be transferred. So molecules with no ionisable sites cannot exist as a salt. For molecules where ionisable sites exist, a rule of thumb related to their pKa exists.34 Accordingly, one looks at the value of ΔpKa = pKa (conjugate acid of base) – pKa (acid). If ΔpKa is less than 0 a co-crystal should be synthesized, whereas a value greater than 2-3 should lead to a salt. Values between 0-3 can predict either case and are dependent on the nature of the components in question. An example of where the pKa rule has been used is the work by Lemmerer et al.35 In this work a series of co-crystals and molecular salts consisting of 2-chloro-4-nitrobenzoic acid with pyridines with varying pKa’s were synthesized, with the aim being to determine if calculated pKa’s were applicable to predicting the formation of a cocrystal or molecular salt. The ΔpKa of each crystal system along with other co-crystals and salts containing 2-chloro-4-nitrobenzoic acid and similar compounds reported in the CSD was calculated. The results showed that most of the salts formed when the ΔpKa was greater than three and most of the cocrystals formed when the ΔpKa was less than zero. Results between zero and three varied. The conclusion of the study shows that the ΔpKa rule seems to work rather well. 1.2.5.2. Pharmaceutical Cocrystals The development of an API from concept to market can take several years and can cost millions of dollars. API’s in development undergo several tests and evaluations to prove their safety and effectiveness, with most failing before being able to reach the market. Therefore, it can be more economically feasible to improve currently existing drugs than to develop new drugs from scratch or modify them. Most API’s exist in the solid form, either in tablet or capsule form. As such it can exist in different solid-state forms. Amorphous API’s have no long-range or 23 periodic order, and typically have higher dissolution rates and higher bioavailability than their more crystalline form.36 However, amorphous forms are undesired because most are unstable, and will convert to a crystalline form anyway. Polymorphs offer the potential of having different physicochemical properties, which can be used to select one form with the best properties. The disadvantage of this is that typically the metastable forms are usually the most desirable since they typically have the best properties among all other known forms. This can prove problematic when the metastable form is unable to be obtained consistently (“disappearing polymorphs”) or converts to a more stable form. Salt formation remains one of the most popular means of improving API’s, with more than half of the API’s available on the market in a salt form.37 The problem with salt formation is that the API must have an ionisable site, which may not be present. Considering all this, without modifying the API, cocrystals can overcome the limitations mentioned above. Cocrystals are becoming increasingly more important due to the different physical properties that can arise. For example, flufenamic acid is a drug reported to have poor water solubility. A co-crystal of flufenamic acid with 2-chloro-4-nitrobenzoic acid was determined to have a solubility of 5 times that of the pure flufenamic acid.38 In some cases, multi-component crystals can include two or more API’s into one solid form. For example, sulfamethoxazole is often used with trimethoprim to treat bacterial infections.39 By involving these two API’s into one solid form, it offers the advantage of combining drugs with different modes of action into one oral dose. The improvement of the physical properties of the API or the inclusion of several APIs into one form can not only allow pharmaceutical companies to offer a superior product but can also save time and money by avoiding repeating clinical trials. It is not possible to predict whether the formation of a cocrystal is guaranteed.40 Several approaches to designing cocrystals exist. One of the most common approaches to cocrystal design is to focus on the use of synthons. Synthons are “structural units within supermolecules 24 which can be formed and/or assembled by known or conceivable synthetic operations involving intermolecular interactions.”41 Many organic compounds with the same functional groups tend to share similar synthons. For example, most primary amides and carboxylic acids tend to form dimers with the 𝑅2 2(8) hydrogen bond motif in the solid state. Synthons made up of identical functional groups are termed as homosynthons, while synthons made up of different functional groups as heterosynthons. One of the best heterosynthons to use for designing cocrystals is that of the pyridine···carboxylic acid pair. An analysis of the CSD showed that this synthon was present in 98% of crystal structures containing said functional groups.42 1.2.7 Tautomerism Tautomerism refers to the ability to convert between two isomers in which only the electrons and/or protons have shifted position. Tautomers are considered to be distinct chemical species, with each form having unique spectra identifying each one. For example ketones can generally undergo tautomerism to form the enol form, although the keto form will be more dominant in general. Different tautomers can be observed in crystal structures, for example, Form VI of sulfapyridine exhibits both the sulfonamide and sulfonimide forms in the same crystal structure.43 Sulfasalazine’s two polymorphs each exhibit different tautomers from each other. Sulfamethazine has been observed to form both the sulfonamide and sulfonimide forms in the same cocrystal with theophylline.44 The tautomer that is favoured may be dependent on the phase the material is in. For example sulfapyridine in the gas phase favours the sulfonamide form,45 while in the solid state the more stable polymorphic forms (I-V) all favoured the sulfonimide form.43,46–48 Tautomerism has gained significant attention over the past years, over the implications of molecular recognition. For example, adenine in its most stable tautomer, pairs with thymine in DNA. However, under irradiation of UV light, adenine undergoes tautomerism. This new tautomer is then mistaken as guanine, which then pairs with cytosine, 25 resulting in a miscoded pair (Fig. 1.2.7).49 For the solid state not only does protic tautomerism offer possible variable synthons for crystal engineering possibilities, it also offers possible applications for photochromism, thermochromism, photochemical hole-burning and hydrogen- bonded dielectrics.49 Therefore, it would be of interest to determine the conditions required to obtain the different tautomers, and exploit them for these various applications. Scheme 1.2.7. The tautomerism of adenine: (A – major tautomer of adenine) hydrogen bonded to thymine (T) on the left versus adenine’s minor tautomer (A’) hydrogen bonded to cytosine (C) on the right. 1.3 Hirshfeld Surfaces The Hirshfeld surface is a method used to visualise the space molecules occupy in the crystal structure. It provides information on various intermolecular interactions occurring in the crystal structure, including weak intermolecular interactions such as weak C-H hydrogen bonding, C- H···π interactions, π-π stacking etc. This method was named in honour of F.L. Hirshfeld after his “stockholder partitioning scheme” but the method itself was developed by McKinnon, Spackman, Mitchell et al. 50 26 The Hirshfeld method is based on partitioning the molecular crystal’s electron density into regions where the electron density is a sum of the spherical atoms of the molecule (the promolecule) that dominates over the sum of the molecule. A weighing function w(r) defined as: 𝑤(𝒓) = ∑ 𝜌𝑎(𝒓) 𝑎 ∈𝑚𝑜𝑙𝑒𝑐𝑢𝑙𝑒 ∑ 𝜌𝑎(𝒓) 𝑎 ∈ 𝑐𝑟𝑦𝑠𝑡𝑎𝑙 ⁄ = 𝜌𝑝𝑟𝑜𝑚𝑜𝑙𝑒𝑐𝑢𝑙𝑒(𝒓) 𝜌𝑝𝑟𝑜𝑐𝑟𝑦𝑠𝑡𝑎𝑙(𝒓)⁄ ≃ 𝜌𝑚𝑜𝑙𝑒𝑐𝑢𝑙𝑒(𝒓) 𝜌𝑐𝑟𝑦𝑠𝑡𝑎𝑙(𝒓)⁄ Where ρ(r) is the spherically averaged Hartree-Fock atomic electron density function. The weighing function is a continuous function in the 0 ≤ w(r) ≤ 1 region. The Hirshfeld surface is defined as where w(r) = 0.5, also called the isosurface. Using this, changes to the molecule will yield changes to the surface. This means that intermolecular interactions such as hydrogen bonding will have an effect on this surface. Different types of surfaces can be defined from the Hirshfeld surface, which can describe different interactions between the molecule(s) and surfaces. Surfaces are colour-coded in different colour ranges depending on the property of interest. The first set is the distance between a molecule inside the surface to the surface di and the distance between the surface to a molecule outside the surface de (Fig. 1.3.1). These two variables are combined together to form dnorm, which is defined as follows: 𝑑𝑛𝑜𝑟𝑚 = 𝑑𝑖 − 𝑟𝑖 𝑣𝑑𝑊 𝑟𝑖 𝑣𝑑𝑊 + 𝑑𝑒 − 𝑟𝑒 𝑣𝑑𝑊 𝑟𝑒 𝑣𝑑𝑊 Where rvdW represents the van der Waals radius of the respective atom of the internal or external molecule. On a colour-code scheme for dnorm, the colours range from red through white to blue. Red spots indicate close contacts while blue spots indicate distant contacts. Observe in Fig. 27 1.3.2 the red spots in both di and de for 4hba. The bright red spot in di is located at the O-H portion of the carboxylate of 4hba while the red spot in the de represent the C=O portion of the carboxylate. The O-H portion represents the hydrogen bond donor and C=O represents the hydrogen bond acceptor. These spots are red indicating short contacts, which is expected since these groups are involved in hydrogen bonding. The dnorm shows both groups to be red since it represents the combination of the di and de. Figure 1.3.1 A plain surface showing the definition of di and de of 4-hydroxybenzoic acid (Refcode JOZZHI01). The di, de and dnorm are not the only important surfaces to consider. The shape index and curvedness are two surface that are also important to consider. Before the shape index and curvedness can be defined, two variables called the principal curvature, κ1 and κ2 must first be defined. These are defined as follows: 𝜅1 = − 1 |𝒏| 𝜕2𝑤 𝜕𝑢2 𝜅2 = − 1 |𝒏| 𝜕2𝑤 𝜕𝑣2 28 Where n is the gradient and u and v are the respective principal directions. From these two equations, the shape index S and curvedness C are defined below: 𝑆 = 2 𝜋 𝑎𝑟𝑐𝑡𝑎𝑛 ( 𝜅1 + 𝜅2 𝜅1 − 𝜅2 ) 𝐶 = 2 𝜋 ln √ 𝜅1 2 + 𝜅2 2 2 The shape index are colour coded from red through to orange, yellow, green to blue. It is a qualitative measure of the “bump” (blue) and “hollows” (red) in a surface. The curvedness indicates regions where the surface is flat. The curvedness and shape index is useful for identifying weaker interactions. π···π stacking for example, can be identified by looking for flat regions in the curvedness surface, while blue and red triangles pointing at each other in the shape index indicates the presence of π···π stacking. 29 Figure 1.3.2 The Hirshfeld surface of 4hba (Refcode JOZZHI01) showing (a) di, (b) de, (c) dnorm, (d) the shape index and (e) the curvedness. Another useful feature of Hirshfeld surfaces are the fingerprint plot. The fingerprint plots are a 2D representation of the different contacts between different atoms to the surface.50 These are made by plotting the di (x-axis) against the de (y-axis), using distance ranges from 0.6 Å to 2.4 Å (Fig. 1.3.3). Visually these plots represent dots, spikes and wings, for example, Fig. 1.3.3. represents the fingerprint plot of 4hba. The spikes with tips at the 1.0 Å mark represent the O···H set of contacts. This is expected since the major intermolecular interaction present in 4hba was hydrogen bonding, which would be composed of short contacts. Longer contacts that form in the > 2.0 Å regions are composed of C···H contacts, which correspond to the weaker C-H hydrogen bonds. Overall, Hirshfeld surfaces prove useful in the analysis of crystal structures, especially for polymorphs and multicomponent crystals. Molecules in different polymorphic forms will adopt either different conformations or exhibit different intermolecular interactions. These can be easily examined by comparing either the Hirshfeld surfaces themselves or the fingerprint plots. 30 Figure 1.3.3 A fingerprint plot of 4hba (Refcode JOZZHI01). 1.4 Sulfa Drugs “And thus, we had the powder of life, born from stone, the sulfa drug.” Dr. Stone, Riichiro Inagaki, Chapter 40 In the early 1920s and 1930s the number of bacterial infections reported across the world was high, and the number of possible treatments were limited. These infections would prove fatal and a demand for a synthetic treatment was born. Sulfanilamide (Scheme 1.4) was first synthesized as early as the 1910s, but its significance was not recognized until much later. In the 1920s, Bayer (a subsidiary of I.G. Farben back then) investigated the use of dyes as a potential source to use as an antibacterial.51 It was assumed that the inclusion of a sulfonamide group within the dye would yield a molecule with anti-bacterial properties. Under Gerhard Domagk, these compounds were tested in vivo of mice. The first successful result of this series of investigations was prontosil, a prodrug which is metabolised in vivo to form sulfanilamide and 5-aminosalicylic acid. Prontosil would later go on to save several thousands of lives and earn Domagk the Nobel Prize for medicine in 1939. It was initially believed that prontosil acted as a bactericidal agent (an agent which directly killed the bacteria) as opposed to being a 31 bacteriostatic agent (an agent which prevented the growth and spread of new bacteria), and the dye portion of the molecule was integral in its anti-bacterial properties. Only much later was it proven that the opposite was true. The demand for sulfa drugs declined after penicillin entered the market as a superior anti-bacterial drug. Nevertheless sulfa drugs are still being used for both human and veterinarian purposes. Scheme 1.4. The basic structure of sulfonanilamide. As the name implies, sulfonamides consist of a sulfonyl group bonded directly to an amine group. For the purpose of this work, only derivatives based closely on sulfonanilamide will be considered. In particular, for this work, the sulfa drugs sulfapyridine (sp), sulfamethazine (sz), sulfamethoxazole (sm) and sulfathiazole (st) were investigated. The mechanism of action of sulfanilamides involves the inhibition of folic acid synthesis, which is a necessary component for the growth of new bacteria. Sulfanilamides are structural analogues of 4-aminobenzoic acid (PABA), an important precursor necessary for folic acid synthesis. Sulfanilamides compete for the dihydropteroate synthetase enzyme with PABA,52 which outcompetes PABA and thus prevents the conversion of PABA to folic acid. 32 1.4.1 Sulfapyridine sp (Scheme 1.4.1) was first reported as an improved treatment for pneumococcal infections over prontosil in 1938 by Lionel Whitby.53 It would later be used in the Second World War to save many lives, including Winston Churchill who fell ill with pneumonia.54 sp consists of the sulfanilamide molecule with one hydrogen from the sulfonamide group replaced with a pyridine group at the 2’ position of the pyridine. sp is polymorphic, consisting of at least six reported forms,43,46,48,55,56 although identifying the forms between works has been inconsistent.48 In all forms sulfapyridine exists in the imidine form, with the exception being form VI where both tautomers exist in the same structure.43 According to Adsmond and Grant, sulfapyridine existing in the sulfonimide form as opposed to the sulfonamide form in the solid state is an exception to the norm.57 Scheme 1.4.1 Basic structure of sulfapyridine showing its tautomerism. Multicomponent crystals containing sp are rare. A few crystal structures of sp solvates containing sp exists.47,58,59 Prior to this work the only cocrystals and molecular salts reported were done by Shunje.60 Shunje reported that sp formed cocrystals with 3-nitrobenzoic acid, dmap and 5-bromosalicylic acid and molecular salts with 4nba, 3,5-dinitrosalicylic acid and 3,5-dibromosalicylic acid. sp showed diversity in the major hydrogen bonds observed. 33 1.4.2. Sulfathiazole st is a sulfanilamide derivative with a 1,3-thiazole ring attached to the sulfonamide group at the 2’ position. It first entered the market in 1940, being developed by several different companies at the time.61 Sulfathiazole is polymorphic, with five forms reported so far. Accordingly, the thermal stability of these polymorphs is in the order of III ≈ IV > II > I, with I being the metastable form.62 Form I also forms a unique dimer different to the other polymorphic forms of sulfathiazole. This is a 𝑅2 2(8) ring hydrogen bond motif, called the α- dimer. While the other forms share a common second level dimer, referred to as the β-dimer. Form I shows no other phase transitions on heating before melting at 201 °C. Forms II-V convert to Form I on heating, before melting at 201 °C. In all cases sulfathiazole exists in the sulfonimide form. One point to note about sulfathiazole is its ability to form solvates. It has been claimed that over one hundred solvates have been reported in the literature,63 but only a fraction of crystal structures of the solvates have been reported.62,64,65 Some notable cocrystals featuring st include forming cocrystals with: theophylline,66 sulfanilamide,66 4nba,67 4-aminobenzamide,68 and glutaric acid.69 In all entries of the CSD containing the neutral form of st, the sulfonimide tautomer of st was reported with no entries reporting the sulfonamide form. This suggests that the sulfonimide form is the most dominant form in the solid state, although no extensive study on the tautomerism of st (either in the solid state or solution) has been done. Scheme 1.4.2 The structure of st showing the sulfonimide tautomer. 34 1.4.3. Sulfamethoxazole sm (Scheme 1.4.3) is a sulfanilamide derivative with a 5-methyl-1,2-oxazole present at the 3’ position. sm is still used today, packaged together with trimethoprim to form co-trimoxazole, and sold under the brand names Bactrim® or Spetrin®. According to the Biopharmaceutics Classification System (BCS), sulfamethoxazole is classified as a class IV drug, which means that it has both low solubility and low permeability. Sulfamethoxazole, like most of the other sulfa drugs, is polymorphic consisting of at least five reported forms.70–73 The first two forms are both monoclinic crystal system C2/c first reported in the 1980s.71,73 The third and fourth forms were obtained using polymer heteronuclei seeding and crystallized in the monoclinic crystal system with space groups P21/c and P21/n, respectively. A fifth polymorph was reported in the 1980s, but crystallographic information regarding it is scarce.70 The most notable multi-component crystal with sulfamethoxazole is the molecular salt consisting of sulfamethoxazole and trimethoprim. First reported as a cocrystal by Giuseppetti and Tandini in 1980,74 it was later rectified as a molecular salt by Nakai et al.75 by use of IR analysis in 1984. This salt consists of an ionic hydrogen bond ring between the pyrimidine ring of trimethoprim and the oxazole/sulfamide group of sulfamthoxazole. Otherwise very few crystal structures of multi-component crystals consisting of sulfamethoxazole have been reported, with other notable examples including coformers with pyridine rings.76 Scheme 1.4.3 Structure of sm. 35 1.4.4. Sulfamethazine sz (Scheme 1.4.4.), also referred to as sulfadimidine, consists of the sulfanilamide with a 4,6- dimethylpyrimidine ring attached. Sulfamethazine was first synthesized in 1941 and used during the 1950s until it was discovered to be carcinogenic.77,78 However, it is still used for veterinarian purposes.78 Unlike most of the other sulfa drugs presented here, only one crystal form of sz has been reported, although it can exist in multiple crystal habits.79 sz has been featured in several crystal engineering studies exhibiting often rare structural features not often observed in other crystal systems. For example, a study done by Perumalla et al. on a system consisting of sulfamethazine and saccharine showed a case of salt-cocrystal polymorphism: a case where one polymorphic form is a cocrystal while the other is a molecular salt.80 Another unusual case is a 2:1 cocrystal containing sz and thp in which both tautomers were present.44 Overall, it seems sz readily forms cocrystals with carboxylic acids.77,81–84 In most of these cases sz adopts the sulfonamide tautomer with the only notable exception being the case of the cocrystal with 4hba, where it adopts the sulfonimide form.82 Scheme 1.4.4. Structure of sz showing the sulfonamide tautomer (right) and sulfonimide tautomer (left). 1.5 Isoniazid 1.5.1 Introduction to isoniazid Isoniazid (inh), an anti-bacterial prodrug that was initially combined with p-aminosalicylic acid (PABA) and streptomycin as a treatment for tuberculosis (TB) in the 1940s.85 inh was 36 later combined with rifampicin, pyrazinamide and ethambutol in a fixed dose concentration tablet that proved to be a more effective treatment towards the late 20th century.85,86 inh is a simple molecule, consisting of a pyridine and a hydrazine group. The mechanism of action involves the conversion of inh to the active form isonicotinic acyl radical by the catalase- peroxidase enzyme KatG present in Mycobacterium tuberculosis bacterium.87 This isonicotinic acyl radical couples with NADH which forms a nicotinoyl-NAD adduct. This complex binds enoyl-acyl carrier protein reductase InhA, which blocks the natural enoyl-AcpM substrate and the action of fatty acid synthase. This ultimately prevents the synthesis of mycolic acid which is a key component of the cell walls of Mycobacterium tuberculosis. Inh has attracted much attention from the crystal engineering community for its simple structure and its effective ability to form cocrystals and molecular salts readily. For a long extent of its history inh was considered to be monomorphic,88–91 with new polymorphic forms only discovered recently.23 Three forms have been reported so far, with single crystal structure data available for two of them. Form I (Z’ = 1) consists of a series of infinitely hydrogen bond chain motifs, formed by the interactions of hydrazide groups. Form II (Z’ = 4) consists of dimers formed between the hydrogen bonded interactions between the oxygen of the amide with one of the hydrogens from the hydrazide group. From this, it can be observed that the hydrazine group dominates most of the hydrogen bond interactions. 1.5.2 Blocking the hydrazine group From a crystal engineering perspective, it may prove to be difficult to work with several hydrogen bond donors/acceptors. Sometimes it may be necessary to remove some of these hydrogen bond donors/acceptors in order to simplify the major hydrogen bonding schemes or the dimensionality of the structure. Previous work by Lemmerer experimented with using 37 “covalent-assisted supramolecular chemistry” which involved using a ketone and reacting it with isoniazid in a Schiff-base reaction. 90,92–94 Here the hydrazine is replaced with an imine, effectively removing two hydrogen bond donors (Scheme 1.5.3). Scheme. 1.5.3. General scheme of the Schiff-base condensation reaction of isoniazid with a ketone or aldehyde (R1 = H). 1.5.3 Polymorphism of isoniazid and its derivatives/cocrystals Some derivatives of inh were also found to be polymorphic, a few of which are described below. One of the most remarkable polymorphic systems of an inh derivative is that of inh reacted with acetophenone, to form (E)-(1-phenylethylidene)hydrazide (IPH), reported by Hean et al.95 Hean and co-workers reported six polymorphic forms for this compound, all of which were obtained using standard methods and materials. These forms were ordered in terms of increasing melting points. All forms, with the exception of form II, form catamers between the amide groups of neighbouring IPH molecules, while form II and the one pair of molecules in the asymmetric unit of form I form a ring hydrogen bond motif. It should be noted that the pyridine ring only forms a strong hydrogen bond in form I. Cocrystals of inh can also exhibit polymorphism, with coformers such as ca (and its derivatives),96–98 PABA,86 4hba hydrate,99 and various simple dicarboxylic acids such as sa, fumaric acid and adipic acid.100 1.6 Aims and Objectives As noted from Chapter 1.4 many of the sulfa drugs discussed exhibit interesting properties such as tautomerism and polymorphism. With the exception of sz, each sulfa drug discussed has at 38 least five polymorphs. Usually polymorphic compounds make good coformers due to the greater structural flexibility.101 Structures of sz, st and sp show the possibility of solid-state tautomerism. Tautomerism in the solid state is still poorly understood, and its understanding depends on finding cases where it exists. Therefore it becomes critical to find these cases. For the sulfa drugs, the number of reported cocrystals and molecular salts is relatively small. The design and synthesis of new crystal forms often requires knowledge of the types of intermolecular interactions that occur. The lack of this knowledge can hamper the design of new solids. Therefore it becomes crucial to explore unknown structural landscapes of these sulfa drugs. 3,5-dinitrobenzoic acid (dnba) is a popular co-former used in crystal engineering studies. At the start of this work, at least 362 entries containg dnba (or its benzoate form) were reported. Despite this, very few studies centred on dnba exists. To my knowledge, there are no other significant applications for this compound, other than being an excellent co-crystalizing agent. Due to the ring of dnba being π-electron poor due to the strong electron withdrawing groups, it makes a potential charge transfer acceptor molecule. Therefore, it would be of interest to explore the multi-component structural landscape of dnba by using a diverse range of coformers. To summarize, the main objectives of this work are to explore and expand the crystal structural landscape of the previously described sulfa drugs, isoniazid derivatives and dnba. In particular:  Design and synthesize new cocrystals and molecular salts of the sulfa drugs sz, st, sm and sp.  Design and synthesize new cocrystals and molecular salts of dnba, and explore the weaker intermolecular interactions using Hirshfeld surfaces. 39  Synthesize and characterize the polymorphs of the diacetone alcohol derivative of isoniazid (iz4h4m2p), as well as its cocrystals. Chapter 2 Experimental Procedures and Aims and Objectives 2.1 Cocrystal/molecular salt synthetic methods Various methods, both solvent-based and mechanochemical-based, have been used to synthesize the multi-component crystals presented in this work. 4,4’-bipyridine was purchased from Alfa-Aeser. All other materials were purchased from Sigma-Aldrich. All solvents were of analytical grade. Table 2.1 lists all the coformers used as part of the cocrystallization attempts in this work. Table 2.1 List of coformers used in this work. Compound name 3-cyanopyridine 4-hydroxypyridine 2-hydroxybenzoic acid (Salicylic acid) 4-cyanopyridine 2-amino-3-nitropyridine 3-hydroxybenoic acid 4-dimethylaminopyridine 2-amino-5-nitropyridine 4-hydroxybenzoic acid 2-aminopyridine 4,4’-bipyridine 3,4-dichlorobenzoic acid 3-aminopyridine 2-amino-5-chloropyridine toluic Acid 4-aminopyridine 3,5-dinitrobenzoic acid saccharine 2-hydroxypyridine 4-nitrobenzoic acid succinic acid 3-hydroxypyridine 2-chloro-4-nitrobenzoic acid cinnamic Acid (trans) 2-chloro-5-nitrobenzoic acid Solution Based Methods Varying stoichiometric amounts of co-former:API (or dnba for chapter 5) were completely dissolved in an appropriate solvent system. In most cases some heating was used to achieve complete dissolution. This solution was then left partially open to allow solvent evaporation for several days, with the aim of leaving crystals suitable for X-ray diffraction behind. Mechanochemical Methods 40 Stoichiometric amounts of the co-formers were placed in a Teflon-coated milling jar (18 mm diameter, 65 mm length) with two steel balls (5 mm diameter). In some cases a few drops of an appropriate solvent was added to the mixture. The jars were then placed on a Retsch MM200 vibratory mill and were milled at 20 Hz for some time (typically between 20-60 minutes). To obtain single crystals suitable for single crystal X-ray diffraction (SC-XRD) from these powders, the same solution based methods described previously were used, where some of the powder obtained from milling was used as seed crystals. 2.2 Synthesis of iz4h4m2p A general method of obtaining iz4h4m2p is given as follows: 0.250 g of isoniazid (1.82 mmol) was dissolved in 10 mL of ethanol (absolute). A few drops (0.5 mL) of 4h4m2p was added to this mixture. This solution was refluxed for 18 h, after which it was transferred to a vial, which was kept loosely open. Crystals were grown from this mixture after several days. Scheme 2.2 Reaction scheme for the synthesis of iz4h4m2p. 2.3 Single Crystal X-Ray Diffraction The Bruker D8 VENTURE PHOTON CMOS area detector diffractometer, equipped with a graphite monochromated Mo-Kα1 sealed tube (50 kV, 30 mA), was used to collect all the intensity data. The program SAINT+, vers. 6.02102 was used to integrate the data and the program SADABS23 was used to make empirical absorption corrections. Space group assignments were made using XPREP102 on all compounds. In all cases, the structures were 41 solved in the WinGX Suite of programs103 by direct methods using SHELXS-97104 and refined using full-matrix least-squares/difference Fourier techniques on F2 using SHELXL-97.105 All non-hydrogen atoms were refined anisotropically. Thereafter, all hydrogen atoms attached to N or O atoms, except where otherwise stated, were located in the difference Fourier map and their coordinates and isotropic thermal displacement parameters were refined freely. Diagrams and publication material were generated using ORTEP-3106 and MERCURY.107 2.4 Powder X-Ray Diffraction Powder X-ray diffraction is done to determine the phase purity of the bulk material. PXRD data for all compounds were measured at 293 K on a Bruker D2 Phaser diffractometer which employs a sealed-tube Co X-ray source (λ = 1.78896 Å), operating at 30 kV and 10 mA, and LynxEye PSD detector in Bragg-Brentano geometry. Powder patterns for each sample are presented in the supporting information, where the experimentally obtained data were compared with the calculated patterns obtained from the SC-XRD data. 2.5 Infrared Spectroscopy FTIR spectra were collected using Bruker Tensor 27 and a Bruker Alpha II model equipped with the Eco-ATR sampling module. Background noise was subtracted and a small amount of the sample of interest was placed onto the ATR crystal. Spectra were measured from 600 to 4000 cm-1 range, resolution 4 cm−1 with 24 scans per sample. The FTIR spectrophotometers were in a room with conditioning at 22 °C. After spectral acquisition the ATR crystal was cleaned using isopropanol. 2.6 Differential Scanning Calorimetry Differential scanning calorimetry data were collected using a Mettler Toledo 822e and a Mettler Toledo DSC 3 with aluminium pans under nitrogen gas (flow rate = 10 mL/min). Exothermic events were shown as peaks. Samples were heated and cooled to determine melting points as 42 well as any additional phase transitions. The temperature and energy calibrations were performed using pure indium (purity 99.99%, m.p. 156.6 °C, heat of fusion: 28.45 J g-1) and pure zinc (purity 99.99%, m.p. 479.5 °C, heat of fusion: 107.5 J g-1). Samples were heated and cooled to different temperatures at different rates. 2.7 Hirshfeld Surface Analysis The Hirshfeld surfaces and fingerprint plots presented were generated using the program Crystal Explorer (vers. 17.5).108 All surfaces were generated using a High (standard) resolution. 2.8 Hot Stage Microscopy Hot stage microscopy was done on an Olympus SZ61 equipped with a Kofler hot bench. Images were captured using an Olympus EP50 camera. Various heating and cooling techniques were applied to capture the phase transitions. 2.9 Thermogravimetric Analysis Thermogravimetric analysis (TGA) was done using a PerkinElmer TGA 6000 thermogravimetric analyzer using high-purity nitrogen at a heating rate of 10 °C/min and gas flow rate of 10 mL/min. 43 Chapter 3 Synthesis and Characterization of a series of Sulfamethazine multi-component crystals with Various Benzoic Acids. 3.1 introduction Nine cocrystals containing sz and a carboxylic acid were synthesized and characterized. These cocrystals were synthesized using both solution and mechanochemical methods. It should be noted that for sz both carboxylic acids and pyridines were used for cocrystallization attempts in the same way as for sp, st and sm. No new crystal structures containing sz and a pyridine derivative was successfully synthesized, which resulted in the theme of the paper being centred on the success of the multi-component crystals containing sz with benzoic acid and its derivatives. Authors: Scheepers, M.C., Lemmerer A. Journal: Crystal Growth & Design Date Published: 3 December 2019 Volume Number: 20 Pages: 813-823 Number of citations as of 10 November 2022: 7 Synthesis and Characterization of a Series of Sulfamethazine Multicomponent Crystals with Various Benzoic Acids Published as part of a Crystal Growth and Design virtual special issue Remembering the Contributions and Life of Prof. Joel Bernstein Matthew C. Scheepers and Andreas Lemmerer* Molecular Sciences Institute, School of Chemistry, University of the Witwatersrand, 2050, Johannesburg, South Africa *S Supporting Information ABSTRACT: Nine multicomponent crystals consisting of sulfamethazine (sz) with benzoic acid and its derivatives were synthesized and characterized. Eight of the nine multicomponent crystals are cocrystals, while one is a molecular salt. The coformers used to form multicom- ponent crystals with sz include 2-chloro-4-nitrobenzoic acid (2c4n), 2-chloro-5-nitrobenzoic acid (2c5n), salicylic acid (2hba), 3-hydroxybenzoic acid (3hba), 4-hydroxybenzoic acid (4hba), 4-bromobenzoic acid (4Brba), benzoic acid (ba), cinnamic acid (ca), and toluic acid (ta). These multicomponent crystals were characterized by single-crystal X-ray diffraction (SC-XRD), powder X-ray diffraction (PXRD), and differential scanning calorimetry (DSC). SC-XRD showed that eight of the coformers that interacted with sz formed the amidine- carboxyl synthon; the only exception to this was sz + 4hba, which formed the imidine-carboxyl synthon formed instead. PXRD confirmed that the single crystals were representative of the bulk material. DSC showed most of the multicomponent crystals to have only a melting phase transition, which differed from the melting points of the coformers. The only exceptions were sz + 4brba and sz + ca, where an additional endothermic peak was observed, which corresponds to an amorphous phase transition before melting. ■ INTRODUCTION The need for designing new crystal forms of potentially useful active pharmaceutical ingredients (APIs) has been rising in the past few decades. Promising drug candidates often have poor physiochemical properties, especially poor water solubility and dissolution rates.1,2 One common approach to altering the poor physiochemical properties of an API is to incorporate them into multicomponent crystals.3 Multicomponent crystals often have different properties, if not superior properties, with respect to its coformers.4 Therefore, there is a great interest in exploring new multicomponent crystals of various APIs. Sulfamethazine (sz), also referred to as sulfadimidine, is an antibacterial drug used to treat urinary tract infection, chlamydia, malaria, rheumatoid fever, and toxoplasmosis in humans as well as for veterinary purposes.5 Sz is also known to be tautomeric,5,6 consisting of an amidine form and an imidine form (Scheme 1). Both forms have been reported to appear in the solid state.5−7 However, the amidine form tends to be the more dominant form, appearing more frequently than the imidine form. In its pure form, sz is able to crystallize into several different crystal habits, but has only one crystal structure.8 In addition to the amidine/ imidine group, an amino group is present to act as an additional hydrogen bond donor, and the sulfoxide and second nitrogen of the pyrimidine ring can act as an additional hydrogen bond acceptor. This allows a wide variety of possible hydrogen bond interactions to occur. Several different cocrystals and solvates of sulfamethazine have already been reported in the literature. The major synthon that forms between sz and the corresponding coformers involves the amidine/imidine functional group. Several multicomponent crystals of sz containing a coformer with a carboxylic acid usually interact with the amidine/imidine group as seen in synthon (i) and (ii) (Scheme 2). In such a case, this should leave the amino and sulfoxide groups free to interact with each other (Synthon (iii), Scheme 2). In this work,we present the synthesis of nine multicomponent crystals of sulfamethazine, with eight cocrystals and one molec- ular salt. Six of these multicomponent crystals are novel, while three have been previously reported but produced under slightly different conditions. The benzoic acids chosen for this work were either structural isomers of each other (2hba, 3hba, 4hba, etc.) or did not contain any additional hydrogen bonding groups Received: September 12, 2019 Revised: December 3, 2019 Published: December 18, 2019 Scheme 1. Sulfamethazine Showing the Amidine Form (Left) and Imidine Form (Right) Article pubs.acs.org/crystalCite This: Cryst. Growth Des. 2020, 20, 813−823 © 2019 American Chemical Society 813 DOI: 10.1021/acs.cgd.9b01209 Cryst. Growth Des. 2020, 20, 813−823 D ow nl oa de d vi a U N IV O F T H E W IT W A T E R SR A N D o n Ju ne 1 3, 2 02 2 at 0 9: 14 :0 7 (U T C ). Se e ht tp s: //p ub s. ac s. or g/ sh ar in gg ui de lin es f or o pt io ns o n ho w to le gi tim at el y sh ar e pu bl is he d ar tic le s. 44 https://pubs.acs.org/page/virtual-collections.html?journal=cgdefu pubs.acs.org/crystal http://pubs.acs.org/action/showCitFormats?doi=10.1021/acs.cgd.9b01209 http://dx.doi.org/10.1021/acs.cgd.9b01209 (e.g., ca, ba, ta). These multicomponent crystals were charac- terized by single crystal X-ray diffraction (SC-XRD), powder X-ray diffraction (PXRD), and differential scanning calorimetry (DSC). ■ EXPERIMENTAL SECTION Crystal Preparation. All reagents used were purchased from Sigma-Aldrich and were used without further purification. Most of the cocrystals presented here have been prepared almost exclusively by solvent evaporation methods. Equimolar amounts of sz and the respective coformer were dissolved together in a 1:1 v/v ratio of ethanol/acetonitrile, which was allowed to stir for an additional 10 min with some heating. Afterward, samples were left slightly open at room temperature to allow solvent evaporation. Crystals were grown from this solution after several days. For the case of sz + 2c5n and sz + ca, ball milling followed by solvent evaporation methods was deployed. Equimolar amounts of sz and the respective coformer were placed in Teflon-coated milling jar (18 mm diameter, 65 mm length) with two steel balls (5 mm diameter). A few drops of acetonitrile were added to this mixture. The jars were then placed on a Retsch MM200 vibratory mill and were milled at 20 Hz for 20 min. Afterward, the solvent evaporation methods described previ- ously were used, with some of the powder from the ball milling used as seeds. Powder X-ray Diffraction (PXRD). Powder X-ray diffraction is done to determine the phase purity of the bulk material. PXRD data for all compounds were measured at 293 K on a Bruker D2 Phaser diffrac- tometer, which employs a sealed tube Co X-ray source (λ = 1.78896 Å), operating at 30 kV and 10 mA, and LynxEye PSD detector in Bragg− Brentano geometry. Powder patterns for each sample are presented in the Supporting Information, where the experimentally obtained data are compared to the calculated patterns obtained from the SC-XRD data. Single Crystal X-ray Diffraction (SC-XRD). The Bruker D8 VENTURE PHOTON CMOS area detector diffractometer, equipped with a graphite monochromated Mo Kα1 sealed tube (50 kV, 30 mA), was used to collect all the intensity data. The program SAINT+, vers. 6.029 was used to integrate the data, and the program SADABS10 was used to make empirical absorption corrections. Space group assignments were made using XPREP10 on all compounds. In all cases, the struc- tures were solved in the WinGX Suite of programs11 by direct methods using SHELXS-9710 and refined using full-matrix least-squares/difference Fourier techniques on F2 using SHELXL-97.10 All non-hydrogen atoms Scheme 2. Some of the Synthons Expected for This Work: (i) Amidine-Carboxyl Synthon, (ii) Imidine-Carboxyl Synthon, (iii) Sulfoxide-Amino Chains, (iv) Amino-Carboxyl Synthon, (v) Hydroxyl-Amino Synthon Scheme 3. Coformers Used for This Work: (i) Cinnamic Acid (ca), (ii) 2-Chloro-4-nitrobenzoic Acid (2c4n), (iii) 2-Chloro- 5-nitrobenzoic acid (2c5n), (iv) 4-Nitrobenzoic Acid (4nba), (v) 4-Bromobenzoic Acid (4brba), (vi) 2-Hydroxybenzoic Acid/Salicylic Acid (2hba), (vii) 3-Hydroxybenzoic Acid (3hba), (viii) 4-Hydroxybenzoic Acid (4hba), (ix) Toluic Acid (ta) Table 1. Grouping of the Multicomponent Crystals As Well As the Motivation for Such Groupings group motivation for grouping together sz + 4Brba sz +4Brba is in its own group since it does not fit in with the other three group list below. sz + ca, sz + ba, sz + ta The coformers consist of a carboxylic acid group which acts as the sole source for hydrogen bonding. sz +2hba, sz +3hba, sz +4hba The coformers are structural isomers of each other. sz +2c4n, sz +2c5n The coformers are structural isomers of each other; in particular, it is the shift of the nitro group from the 4′ position to the 5′ position. Crystal Growth & Design Article DOI: 10.1021/acs.cgd.9b01209 Cryst. Growth Des. 2020, 20, 813−823 814 45 http://pubs.acs.org/doi/suppl/10.1021/acs.cgd.9b01209/suppl_file/cg9b01209_si_001.pdf http://dx.doi.org/10.1021/acs.cgd.9b01209 Figure 1.Ortep diagram for (i) sz + 2c4n, (ii) sz + 2c5n, (iii) sz + 2hba, (iv) sz + 3hba, (v) sz + 4brba, (vi) sz + 4hba, (vii) sz + ba, and (viii) sz + ca. Crystal Growth & Design Article DOI: 10.1021/acs.cgd.9b01209 Cryst. Growth Des. 2020, 20, 813−823 815 46 http://dx.doi.org/10.1021/acs.cgd.9b01209 were refined anisotropically. Thereafter, all hydrogen atoms attached to N or O atoms, except where otherwise stated, were located in the difference Fourier map, and their coordinates and isotropic thermal displacement parameters were refined freely. All C−H hydrogen atoms were placed at idealized positions and refined as riding atoms with isotropic parameters 1.2 or 1.5 times those of the “heavy” atoms to which they are attached. Diagrams and publication material were generated using ORTEP-311 and Mercury.12 Differential Scanning Calorimetry (DSC). Differential scanning calorimetry data were collected using a Mettler Toledo 822e with aluminum pans under nitrogen gas (flow rate = 10 L/min). Exothermic events were shown as peaks. Cocrystals were heated and cooled to determine melting points as well as any additional phase transitions. The temperature and energy calibrations were performed using pure indium (purity 99.99%, mp 156.6 °C, heat of fusion: 28.45 J g−1). Infrared Spectroscopy. Infrared spectra were collected on a Bruker Tensor 27. Samples were used as is without further processing. Spectra were collected in the range of 4000−600 cm−1, and band intensities were expressed as transmittance %. ■ RESULTS AND DISCUSSION Nine multicomponent crystal forms containing sulfamethazine are reported. Six of these crystal forms are new. These crystal forms were characterized by SC-XRD. PXRD was used to con- firm that the bulk material matched the calculated pattern in order to ensure purity. DSC was used to determine the melting points as well as any other potential phase transitions. SC-XRD Results. Sz formed nine multicomponent crystals which were characterized by SC-XRD. For most of these crystal systems, the intermolecular interactions are similar and possess some structural and packing similarities. For the purpose of this work, several multicomponent crystals are grouped and discussed together below. Table 1 gives a list of which multicomponent crystals are grouped together and reasons for such grouping. In addition, Table 4 contains a list of the bond lengths of the carboxyl group obtained from each of the multicomponent crystals as well as if a salt or cocrystal was formed. The ratios of the C−O bond lengths indicate that most of the crystals systems are cocrystals, with the exception of sz + 2c4n and sz + 3hba, which show incomplete proton transfer. Crystal Structures of sz + 4brba. The cocrystal of sz + 4brba crystallizes in the P21/n space group with the asymmetric unit consisting of one molecule of each sz and 4brBA. Sz forms in the amidine tautomer, forming the expected amidine-carboxyl synthon. Sz adopts a v-conformation, which forms wavelike col- umns along the c-axis (Figure 3(i)), with 4brba sitting in between adjacent columns. However, the pyrimidine rings of sz and 4brba forms a layered structure, with layers separated from each other via the aniline ring of sz. Crystal Structures of sz + ba, sz + ca, and sz + ta. The crystal structure of sz + ba has been previously reported,22 while sz + ca and sz + ta are novel to this work. Cocrystals of sz + ba, sz + ca, and sz + ta are described and compared together since the coformers only consist of one hydrogen bond donor/acceptor− the carboxyl group. From this, it should be reasoned that differ- ences in the crystal structures should arise due to the unique aspects of the respective coformer. Cocrystals of sz + ba, sz + ca, and sz + ta crystallize in the Pbca, P21/n, and P21 space groups, respectively. In each case, the asymmetric unit is made up of one molecule of sz and onemolecule of the respective coformer. The coformers ta, ba, and ca contain only one functional group that can form hydrogen bondsthe carboxyl group. In each case, the amidine-carboxyl synthon (Synthon (ii), see Scheme 2) formed. Sz + ca and sz + ta formed the sulfoxide-amino chain synthon (Synthon (iii), see Scheme 2), while sz + ba did not. Instead, in sz + ba the amino group forms a hydrogen bond to one of the oxygen atoms of the sulfoxide group, while the other hydrogen bonds to the carboxyl group of ba (Synthon (iv) see Scheme 2). In the packing of the cocrystal of sz + ba (Figure 4(ii)), the pyrimidine ring of sz and rings of ba form rows of the aniline ring of sz in the b-axis direction separating clusters of sz + ba apart. Adjacent dimers of sz + ba running in the a-axis appear T-shaped, with rows interlocking together. In the packing of sz + ca, molecules of sz lie in rows down the b-axis, alternating between the aniline and pyrimidine rings. This results in the szmolecules forming parallel zigzag formations, with molecules of ca. filling in the gaps between the sz-rows. Crystal Structures of sz + Hydroxybenzoic Acids. The cocrystals of sz + 2hba and sz + 3hba both crystallize in the Pbca space group, while the cocrystal of sz + 4hba crystallizes in the P1̅ space group. In each case, the asymmetric unit is composed of one molecule of sz and each respective conformer. The crystal structures of sz + 4hba and sz + 2hbahave been previously reported in the literature.5,23 Despite using slightly different crystallizing conditions (sz + 2hba used acetone as the solvent, while sz + 4hba used pure acetonitrile), the same crystal structures were Figure 2. Ortep diagram for sz + ta. Crystal Growth & Design Article DOI: 10.1021/acs.cgd.9b01209 Cryst. Growth Des. 2020, 20, 813−823 816 47 http://dx.doi.org/10.1021/acs.cgd.9b01209 obtained. The cocrystal of sz + 3hba is novel; hence, the syn- thesis of the former proved useful in comparing the changes that may occur due to the shift of the hydroxyl group. In the cocrystal of sz + 3hba, it was observed that the carboxylic acid proton had only partially transferred to the sz. In the difference Fourier map, there were two clear areas of electron density near O3 and N4. Subsequently, we decided to place a hydroxyl (H3A) and aro- matic (H4) hydrogen atom geometrically and give each one a site occupancy of a half. This behavior is reflected in the ratio of the C−O bonds given in Table 3. For a typical carboxylic acid, the ratio should be greater than 1.5. For a carboxylate, the value should be close to 1.0. In this case, the ratio was found to be 1.027. Hence, we will refer to sz + 3hba as a mixed salt/cocrystal. sz + 2hba and sz + 3hba form the amidine-carboxyl synthon, while sz + 4hba formed the imidine-carboxyl synthon instead. sz + 2hba and sz + 3hba form the sulfoxide-amino chains, while Sz + 4hba do not. Instead, only one of the oxygen atoms of the sulfoxide group forms a hydrogen bond with the amino group (N1−H1A···O2). The hydroxyl group for sz + 2hba forms an intramolecular hydrogen bond as expected. For both sz + 3hba and sz + 4hba, the hydroxyl group forms a hydrogen bond inter- action with the amino group of sz (Synthon (v), see Scheme 2). The packing of the crystal structures of sz + 2hba and sz + 3hba are isomorphous, as seen in Figure 5(i) and (ii). Like many of the other crystal structures reported in this work, sz adopts a v-shaped conformation. Both sz + 2hba and sz + 3hba form long chains of sz connected by the sulfoxide-amino hydrogen bond, which essentially forms long columns with the hydroxybenzoic acid attached. These columns overall appear in a wavelike structure. The packing of sz + 4hba deviates from the previous two. Two molecules of each sz and 4hba form tetramers, which stack together to form a ladder. This ladder is made up of the pyrimi- dine rings of sz and 4hba making up the steps, while the aniline Figure 4.Crystal structure of sz + ba showing the hydrogen bonding patterns (i) and its packing (ii). The packing diagrams for sz + ca (iii) and sz + ta (iv). Figure 3.Crystal structure of sz + 4brba showing (i) the packing viewed down the b-axis, (ii) the layers of the pyrimidine of sz and 4brba separated by the aniline rings. Crystal Growth & Design Article DOI: 10.1021/acs.cgd.9b01209 Cryst. Growth Des. 2020, 20, 813−823 817 48 http://dx.doi.org/10.1021/acs.cgd.9b01209 ring makes up the rail. This description matches the one previ- ously reported.5 Crystal Structures of sz + 2c4n and sz + 2c5n.Themulti- component crystals of sz + 2c4n and sz + 2c5n form amolecular salt and a cocrystal, respectively. Sz + 2c4n and sz + 2c5n crystallize in the Pca21 and P21 space groups, respectively. The asymmetric unit for sz + 2c4n consists of one molecule of each sz and 2c4n, while the one for sz + 2c5n consists of two mole- cules of sz and 2c5n. As was the case for sz + 3hba, sz + 2c4nwas observed to have a positional disorder of the carboxylic H atom. Table 2. Crystallographic Data and Structure Refinement Data for sz Multicomponent Crystals sz + 2c4n sz + 2c5n sz + 2hba sz + 3hba sz + 4hba empirical formula C19H18ClN5O6S C19H18ClN5O6S C19H20N4O5S C19H20N4O5S C19H20N4O5S formula weight/g·mol−1 479.89 479.89 416.45 416.45 416.45 temperature/K 173(2) 173(2) 173(2) 173(2) 173(2) wavelength/Å 0.71073 0.71073 0.71073 0.71073 0.71073 crystal system orthorhombic monoclinic orthorhombic orthorhombic triclinic space group Pbc21 P21 Pbca Pbca P1̅ a/Å 19.2999(6) 6.96450(10) 10.1289(3) 9.9985(3) 7.9638(9) b/Å 7.7198(2) 19.3894(3) 15.7422(4) 15.7375(4) 9.3601(10) c/Å 13.9038(4) 15.7238(2) 25.0535(7) 24.6844(7) 13.2882(14) α/° 90 90 90 90 74.123(4) β/° 90 90.7000(10) 90 90 75.616(4) γ/° 90 90 90 90 85.727(4) V/Å3 2071.55(10) 2123.14(5) 3994.81(19) 3884.12(19) 922.85(18) Z 4 4 8 8 2 density (calculated)/ Mg m−3 1.539 1.501 1.385 1.424 1.499 μ/mm−1 0.335 0.326 0.201 0.207 0.218 F(000) 992 992 1744 1744 436 crystal size/mm3 0.214 × 0.248 × 0.317 0.349 × 0.189 × 0.130 0.441 × 0.284 × 0.238 0.315 × 0.257 × 0.129 0.163 × 0.146 × 0.052 theta range for data/° collection 2.110 to 28.238 1.295 to 28.215 3.057 to 28.334 2.924 to 28.360 3.126 to 28.334 reflections collected 24549 58621 58505 47634 36343 independent reflections 5046 [R(int) = 0.0545] 10439 [R(int) = 0.0482] 4953 [R(int) = 0.0405] 4831 [R(int) = 0.0748] 4592 [R(int) = 0.0511] goodness-of-fit on F2 1.030 1.059 1.021 1.072 1.046 final R indices [I > 2σ(I)] R1 = 0.0366 R1 = 0.0392 R1 = 0.0373 R1 = 0.0452 R1 = 0.0416 wR2 = 0.0707 wR2 = 0.0854 wR2 = 0.0926 wR2 = 0.1134 wR2 = 0.0977 R indices (all data) R1 = 0.0560 R1 = 0.0595 R1 = 0.0484 R1 = 0.0693 R1 = 0.0555 wR2 = 0.0785 wR2 = 0.0935 wR2 = 0.0999 wR2 = 0.1250 wR2 = 0.1039 CCDC no. 1952592 1952593 1952594 1952595 1952596 sz + 4brba sz + ba sz + ca sz + ta empirical formula C19H19BrN4O4S C19H20N4O4S C21H22N4O4S C20H22N4O4S formula weight/g·mol−1 479.35 400.45 426.48 414.47 temperature/K 173(2) 173(2) 173(2) 173(2) wavelength/Å 0.71073 0.71073 0.71073 0.71073 crystal system monoclinic orthorhombic monoclinic monoclinic space group P21/n Pbca P21/n P21 a/Å 8.5979(7) 15.1817(3) 8.3614(2) 7.3454(6) b/Å 11.8215(10) 14.0288(3) 14.6091(4) 13.3001(11) c/Å 19.5587(16) 17.9844(4) 16.9318(5) 11.2172(10) α/° 90 90 90 90 β/° 102.009(3) 90 98.2910(10) 107.585(3) γ/° 90 90 90 90 V/Å3 1944.4(3) 3830.34(14) 2046.65(10) 1044.65(15) Z 4 8 4 2 density (calculated)/Mg m−3 1.637 1.389 1.384 1.318 μ/mm−1 2.257 0.203 0.195 0.188 F(000) 976 1680 896 436 crystal size/mm3 0.365 × 0.086 × 0.066 0.347 × 0.248 × 0.160 0.514 × 0.317 × 0.254 0.325 × 0.258 × 0.136 theta range for data collection 2.983 to 28.356° 2.265 to 30.749° 2.899 to 28.315° 2.909 to 28.356° reflections collected 26446 46609 40194 19104 independent reflections 4840 [R(int) = 0.0864] 5970 [R(int) = 0.0892] 5084 [R(int) = 0.0413] 5182 [R(int) = 0.0224] goodness-of-fit on F2 1.050 1.011 1.081 1.064 final R indices [I > 2σ(I)] R1 = 0.0485 R1 = 0.0465 R1 = 0.0372 R1 = 0.0336 wR2 = 0.0748 wR2 = 0.1101 wR2 = 0.0955 wR2 = 0.0932 R indices (all data) R1 = 0.0889 R1 = 0.0917 R1 = 0.0478 R1 = 0.0375 wR2 = 0.0840 wR2 = 0.1314 wR2 = 0.1011 wR2 = 0.1025 CCDC no. 1952597 1952598 1952599 1952600 Crystal Growth & Design Article DOI: 10.1021/acs.