View PDF VersionPrevious ArticleNext Article
 
DOI: 10.1039/D4CE00551A (Paper) CrystEngComm, 2024, 26, 5699-5715

Single-step synthesis of multicomponent cocrystals and salts: the role of laboratory seeding†

Sampurna Nayak and Ashwini K. Nangia *
School of Chemistry, University of Hyderabad, Prof. C. R. Rao Road, Gachibowli, Central University P.O., Hyderabad 500 046, India. E-mail: ashwini.nangia@gmail.com; ansc@uohyd.ac.in

Received 1st June 2024 , Accepted 24th August 2024

First published on 27th August 2024

Abstract

The synthesis of multicomponent cocrystals and salts (MCCS) is an active and contemporary theme in the crystal engineering of pharmaceutical crystal forms. The self-assembly process of multiple molecular components to an ordered and organized crystalline structure becomes increasingly difficult and complex as more than two molecules are present, e.g., ternary, quaternary, and higher-order cocrystals (HOC). One of the most frequent synthetic pathways to promote the self-assembly is mechanochemistry, assisted by a small amount of solvent added. Herein we report a comparative study of two mechanochemical routes on a series of ternary drug systems, halogen bonded ternary cocrystals and quaternary molecular cocrystals: (1) the sequential addition of components in different orders, referred to as the M1 method. The nth component is added to the pre-formed adduct of n − 1 remaining components. (2) The addition of all the components (n) in a single step, referred to as the M2 method. In both methods solvent drop grinding (SDG) or liquid-assisted grinding (LAG) manual operation were used. An excellent match of the experimental PXRD pattern from M1 and M2 procedures with that of the calculated PXRD pattern calculated from single-crystal X-ray diffraction (SC-XRD) data was noted for a majority of the systems tested. The role of crystal seeds/nuclei in forming the multicomponent crystalline products was established under optimal conditions in a single-step synthetic protocol. Concurring positive results on quaternary systems validate the applicability of this method to HOCs. When in a few cases complete transformation to the ternary or quaternary phase was not observed after manual grinding, in such systems ball-mill grinding (BMG) proved to complete the cocrystallization. A simple, efficient and scalable grinding method for MCCS is reported which can be extended to higher-order multicomponent cocrystals.

Introduction

The supramolecular synthesis of binary and ternary cocrystals of active pharmaceutical ingredients (APIs) is an important activity in drug discovery and development research. In a recent study with acefylline,1 we noted that the ternary products with different coformers could be prepared by either grinding the two components (API + coformer 1) to provide cocrystal 1 (CC1), and then adding the second coformer 2 (CF2) to the pre-formed CC1 and further grinding afforded crystalline CC2. Similar results were obtained when CF2 was ground first with the API and then CF1 was added to afford the same ternary product CC2 (Scheme 1). Liquid-assisted grinding (LAG)/solvent-drop grinding (SDG) in a mortar and pestle by manual grinding (MPMG) were used.2 In subsequent experiments, we discovered that once the products were isolated and crystallized in our lab, then a single-step grinding of API + CF1 + CF2 also gave the same ternary product CC2. However, this single-step mixing and grinding were not successful in producing CC2 in the early stages of the program. Typically, CC1 or a mixture of crystalline products were observed in the powder X-ray diffraction pattern (PXRD). This led us to believe and hypothesize that seeding in the laboratory environment by the crystal nuclei3 after the initial experiments were carried out could be a reason for the success of the single-step method for the self-organization of ternary components in a crystalline lattice.
image file: d4ce00551a-s1.tif
Scheme 1 The two liquid-assisted/solvent drop grinding methods reported in this paper.

The synthesis of binary cocrystals and salts by grinding and solution crystallization is relatively common, and numerous studies are reported (see a recent review4). However, the supramolecular synthesis becomes more complex5,6 and multiple pathways are possible when three (and more) chemical species are present in the same cocrystallization experiment. Desiraju7,8 proposed Aufbau build-up of complex supramolecular modules from simpler molecules via binary and then ternary and quaternary cocrystals. The presence of molecular clusters and supramolecular synthon bonded modules during self-assembly to cocrystals has been shown by computations and NMR spectroscopy.9–11 On the other hand, another recent study showed that binary phases do form but do not act as intermediates in the formation of ternary cocrystals by mechanochemical milling.12 Mechanochemical synthesis13 of higher-order multicomponent crystals is a complex process and requires a better understanding of reaction kinetics and mechanistic pathways.14,15 The role of crystal seeds and nuclei in directing the course of crystallization is well documented for single-component crystals (in polymorphism).3,16 Recent reports extend the significance of crystal nuclei and seeds even for multicomponent cocrystals.17–20 A number of factors affect the cocrystallization products in mechanochemical milling and grinding,21–24 a process which is also important in pharmaceutical materials and manufacturing.25–31

Multi-drug and drug–coformer–drug cocrystals32 are ternary in nature and other recent examples of ternary and higher-order cocrystals33–35 are the backdrop of the present report. We show in this paper a general LAG/SDG mortar and pestle manual grinding (MPMG) methodology (Scheme 1) for ternary cocrystals and salts of APIs to afford MCCS by a simple and inexpensive green approach with minimal use of solvents. An extensive comparative study of sequential addition and direct addition mechanochemical procedure for multicomponent solid synthesis on a wide range of higher-order cocrystals was implemented. This study contributes to the sequential supramolecular synthesis of HOCs (ternary and quaternary) from lower-component binary systems in the category of pharmaceutical cocrystals and salts. Most importantly, it highlights the significance of crystal seeds and nuclei in ternary and quaternary cocrystal synthesis and further deepens the knowledge of their role. While their impacts are well documented for binary cocrystals, it is lesser known in the area of supramolecular self-assembly of more than two molecular species. They exhibit substantial dominance in making the one-step synthesis reproducible. Such optimization of the single-step direct addition procedure is necessary as it comes with benefits of being less time-consuming and environmentally friendly due to less solvent requirement (compared to sequential addition), which are crucial considerations for scaling up. Neat grinding (NG), namely grinding without solvent addition, was not effective to produce binary/ternary cocrystals and hence was not pursued.

Results and discussion

The ternary systems studied are summarized in Table 1. They are a continuation of the acefylline study1 from our group and some of the preliminary results on ACF cocrystals by grinding were described in a previous report.1 However, the standardization and reproducibility of the two grinding protocols across a variety of cocrystal/salt systems remained to be demonstrated, and that is the scope of this paper. For this purpose, we added a few ternary systems from our own lab36–38 and a few systems from other researchers' labs39,40 which are reported in the literature. The ternary system is defined based on the solid components present in the crystal structure; adventitious inclusion of water during grinding was not considered as a fourth component (entry 1). After analyzing the ternary systems, we picked two quaternary systems41 (Table 2). The number of possible addition modes in the 4-component system is very large (12 possible addition and grinding experiments) in the stepwise, sequential addition method, and hence they are exemplified on only two systems to show that the manual grinding method has the potential for more complex multicomponent cocrystal systems. In cases where the results of mortar and pestle manual grinding (MPMG) were not matching or inconsistent, the solid mixture of components was subjected to ball-mill grinding (BMG) to complete the transformation. Such a mechanochemical grinding gave matching results in ternary and quaternary systems. The “sequence of addition of the molecular components” in the grinding methodology is summarized in Scheme 2 as well as in Tables 3 and 4 (including the solvent used for LAG/BMG), and details are provided in the Experimental section.
Table 1 Examples of ternary cocrystals and salts prepared by stepwise and direct method of mortar and pestle manual grinding (MPMG) and ball-mill grinding (BMG)
S. no. Ternary system Chemical name Cocrystal or salt Ref.
a Inferred cocrystal/salt state from the IR frequency of the carboxylic acid C[double bond, length as m-dash]O group in Table 3 of ref. 38.
1 INH–NA–FA (1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) Isoniazid, nicotinamide, fumaric acid Cocrystal S. Aitipamula et al.40
2 INH–NA–SA (1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) Isoniazid, nicotinamide, succinic acid Cocrystal S. Aitipamula et al.40
3 4NBA–OA–DIB (2[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) 4-Nitrobenzamide, oxalic acid, p-diiodobenzene Cocrystal S. Tothadi et al.39
4 4NBA–FA–DIB (2[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) 4-Nitrobenzamide, fumaric acid, p-diiodobenzene Cocrystal S. Tothadi et al.39
5 4NBA–OA–BIB (2[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) 4-Nitrobenzamide, oxalic acid, p-bromoiodobenzene Cocrystal S. Tothadi et al.39
6 SMBA–NAM–2HP (1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) p-Sulfonamide benzoic acid, nicotinamide, 2-hydroxypyridone Cocrystal G. Bolla et al.37
7 SMBA–NAM–MeHP (1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) p-Sulfonamide benzoic acid, nicotinamide, 6-methyl-2-hydroxypyridone Cocrystal G. Bolla et al.37
8 SMBA–INA–2HP (1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) p-Sulfonamide benzoic acid, isonicotinamide, 2-hydroxypyridone Cocrystal G. Bolla et al.37
9 SMBA–PAM–MeHP (1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) p-Sulfonamide benzoic acid, picolinamide, 6-methyl-2-hydroxypyridone Cocrystal G. Bolla et al.37
10 ACZ–NAM–2HP (1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) Acetazolamide, nicotinamide, 2-hydroxypyridone Cocrystal G. Bolla et al.36
11 BUM–2HP–PCA (1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) Bumetanide, 2-pyridone, 2-picolinic acid Salta S. Allu et al.38
12 BUM–INA–PCA (1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) Bumetanide, isonicotinamide, picolinic acid Salta S. Allu et al.38
13 BUM–INA–PASA (1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) Bumetanide, isonicotinamide, p-aminosalicylic acid Salt-cocrystala S. Allu et al.38
14 BUM–INA–VLA (1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) Bumetanide, isonicotinamide, vanillic acid Salt S. Allu et al.38
15 PRZ–FLA–PRA (1[thin space (1/6-em)]:[thin space (1/6-em)]2[thin space (1/6-em)]:[thin space (1/6-em)]2) Piperazine, ferulic acid, pyrazinamide Cocrystal X.-Z. Yu et al.42
16 ACF–PIP–GA–H2O (1[thin space (1/6-em)]:[thin space (1/6-em)]0.5[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]2) Acefylline, piperazine, gallic acid Salt-cocrystal hydrate S. Allu et al.1
17 ACF–INA–GA (1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) form 1 Acefylline, isonicotinamide, gallic acid Salt-cocrystal S. Allu et al.1

Table 2 Examples of quaternary cocrystals prepared by stepwise and direct methods of grinding (MG, BMG)
S. no. Quaternary system Chemical name Cocrystal or salt Ref.
1 BDHB–TMP–AZO–PYR (2[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) 4-Bromo-3,5-dihydroxybenzoic acid, 2,3,5,6-tetramethylpyrazine, 4,4′-azopyridine, pyrene Cocrystal S. Roy et al.41
2 BDHB–TMP–DPE-II–PYR (2[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) 4-Bromo-3,5-dihydroxybenzoic acid, 2,3,5,6-tetramethylpyrazine, 1,2-bis(4-pyridyl)ethylene, pyrene Cocrystal S. Roy et al.41


image file: d4ce00551a-s2.tif
Scheme 2 LAG/SDG grinding methodology and naming/numbering system followed in this paper. (a) M2 direct addition method for quaternary system; (b) M2 direct addition method for ternary system; (c) for quaternary cocrystals, all possible M1 methods are subdivided into 12 sequential additions based on the order of coformer addition; (d) for ternary cocrystals, M1 methods are subdivided into 3 sequential additions based on the order of coformer addition without API; (e) for ternary cocrystals, M1 methods are subdivided into 2 sequential additions based on the order of coformer addition with API (the API is always added in the first step).
Table 3 Ternary cocrystals and salts and the experimental procedure followed
S. no. Ternary system Method and sequence of addition Figure number and results
a There is a slight shift in the position of peaks for the experimental PXRD and the calculated XRD lines (red plots in overlay figures) for several crystalline solids, but this is a temperature effect and is ignored so long as the shift is systematic and the overall peak profile looks similar. b There is no X-ray crystal structure reported and so calculated XRD lines are not available to compare with. c The peak intensity is much lower for the experimental PXRD compared to the calculated XRD lines due to the low crystallinity of the powder sample after grinding.
1 ACF–PIP–GA–H2O (1[thin space (1/6-em)]:[thin space (1/6-em)]0.5[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]2) M1(a): ACF + PIP, LAG for 30 min. Then add GA, and continue LAG for 30 min Fig. 1
M1(b): ACF + GA, LAG for 30 min. Then add PIP, and continue LAG for 30 min Match between M1(a), M1(b) and M2 PXRD lines
M2: direct addition of ACF + PIP + GA, LAG for 30 min Match with calculated PXRD patterna
MeCN for LAG
2 ACF–INA–GA (1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) form 1 M1(a): ACF + INA, LAG for 30 min. Then add GA, and continue LAG for 30 min Fig. 2
M1(b): ACF + GA, LAG for 30 min. Then add INA, and continue LAG for 30 min Match between M1(a), M1(b) and M2 PXRD lines
M2: direct addition of ACF + INA + GA, LAG for 30 min Match with calculated PXRD patterna
MeCN for LAG
3(a) ACZ–NAM–2HP (1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) M1(a): ACZ + NAM, LAG for 30 min. Then add 2HP, and continue LAG for 30 min Fig. S1†
M1(b): ACZ + 2HP, LAG for 30 min. Then add NAM, and continue LAG for 30 min Match between M1(a), M1(b) and M2 PXRD lines
M2: direct addition of ACZ + NAM + 2HP, LAG for 30 min Mismatch with calculated PXRD pattern
EtOAc for LAG NMR is shown in Fig. S12(a)†
3(b) ACZ–NAM–2HP (1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) M1(a): ACZ + NAM, BMG for 30 min. Then add 2HP, and continue BMG for 30 min Fig. 3
M1(b): ACZ + 2HP, BMG for 30 min. Then add NAM, and continue BMG for 30 min Match between M1(a), M1(b) and M2
M2: direct addition of ACZ + NAM + 2HP, BMG for 30 min Match with calculated PXRD patterna
EtOAc for BMG; milling frequency 25 Hz NMR is shown in Fig. S12(b)†
4 SMBA–NAM–2HP (1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) M1(a): SMBA + NAM, LAG for 30 min. Then add 2HP, and continue LAG for 30 min Fig. 4
M1(b): SMBA + 2HP, LAG for 30 min. Then add NAM, and continue LAG for 30 min Match between M1(a), M1(b) and M2 PXRD lines
M2: direct addition of SMBA + NAM + 2HP, LAG for 30 min Match with calculated PXRD patterna
EtOAc for LAG
5a SMBA–INA–2HP (1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) M1(a): SMBA + INA, LAG for 30 min. Then add 2HP, and continue LAG for 30 min Fig. 5a
M1(b): SMBA + 2HP, LAG for 30 min. Then add INA, and continue LAG for 30 min Match between M1(a), M1(b) and M2 PXRD lines
M2: direct addition of SMBA + INA + 2HP, LAG for 30 min Mismatch with calculated PXRD pattern
EtOAc for LAG Fig. 5b
Expected ternary cocrystal matches calculated PXRD line patterns with binary SMBA–INA (1[thin space (1/6-em)]:[thin space (1/6-em)]1); ternary is absent
1H NMR spectrum is shown in Fig. S13(a) and (b)†
5b SMBA–INA–2HP (1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) M2: direct addition of SMBA + INA + 2HP, BMG for 90 min Fig. 5c
EtOAc for BMG Manual LAG and BMG PXRD XRD line patterns are different from each other and the calculated XRD of ternary cocrystal
1H NMR spectrum is shown in Fig. S13(c) and (d)†
6 BUM–2HP–PCA (1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) M1(a): BUM + 2HP, LAG for 30 min. Then add PCA, and continue LAG for 30 min Fig. 6
M1(b): BUM + PCA, LAG for 30 min. Then add 2HP, and continue LAG for 30 minutes Match between M1(a), M1(b) and M2 PXRD lines
M2: direct addition of BUM + 2HP + PCA, LAG for 30 min PXRD pattern matches with the reported line pattern in the original paperb
EtOAc for LAG 1H NMR spectrum is shown in Fig. S14†
7 BUM–INA–PASA (1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) M1(a): BUM + INA, LAG for 30 min. Then add PASA, and continue LAG for 30 min Fig. 7
M1(b): BUM + PASA, LAG for 30 min. Then add INA, and continue LAG for 30 min Match between M1(a), M1(b) and M2 PXRD lines
M2: direct addition of BUM + INA + PASA, LAG for 30 min PXRD pattern matches with the reported line pattern in the original paperb
EtOAc for LAG 1H NMR spectrum is shown in Fig. S15†
8 4NBA–FA–DIB (2[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) M1(a): 4NBA + FA, LAG for 30 min. Then add DIB, and continue LAG for 30 min Fig. 8
M1(b): 4NBA + DIB, LAG for 30 min. Then add FA, and continue LAG for 30 min Match between M1(a), M1(b), M1(c) and M2 PXRD lines
M1(c): FA + DIB, LAG for 30 min. Then add 4NBA, and continue LAG for 30 min Match with calculated PXRD patterna,c
M2: direct addition of 4NBA + FA + DIB, LAG for 30 min See also entry 6 and 7 in Table S1†
EtOH for LAG
9 INH–NA–SA (1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) M1(a): INH + NA, LAG for 30 min. Then add SA, and continue LAG for 30 min Fig. 9
M1(b): INH + SA, LAG for 30 min. Then add NA, and continue LAG for 30 min Match between M1(a), M1(b), and M2
M2: direct addition of INH + NA + SA, LAG for 30 min Match with calculated PXRD patterna
MeOH for LAG
10 INH–NA–FA (1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) M1(a): INH + NA, LAG for 30 min. Then add FA, and continue LAG for 30 min Fig. 10
M1(b): INH + FA, LAG for 30 min. Then add NA, and continue LAG for 30 min Match between M1(a), M1(b), and M2
M2: direct addition of INH + NA + FA, LAG for 30 min Match with calculated PXRD patterna
MeOH for LAG

Table 4 Quaternary cocrystals and their detailed experimental procedure
S. no. Ternary system Method and sequence of addition Results
a There is a slight shift in the position of peaks for the experimental PXRD and the calculated XRD lines (red plots in overlay figures) for several crystalline solids, but this is a temperature effect and is ignored so long as the shift is systematic and the overall peak profile looks similar.
1 BDHB–TMP–AZO–PYR (2[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) M1(a): BDHB + TMP, LAG for 30 min. Add PYR and continue LAG for 30 min. Add AZO and continue LAG for 30 min Fig. 11
M1(b): BDHB + TMP, LAG for 30 min. Add AZO and continue LAG for 30 min. Add PYR and continue LAG for 30 min Match between M2, M1(a), M1(b), M1(c), M1(d), M1(e), M1(f), M1(k) and M1(l)
M1(c): BDHB + PYR, LAG for 30 min. Add TMP and continue LAG for 30 min. Add AZO and continue LAG for 30 min Match with calculated PXRD patterna and manual grinding in mortar and pestle
M1(d): BDHB + PYR, LAG for 30 min. Add AZO and continue LAG for 30 min. Add TMP and continue LAG for 30 min Fig. 12
M1(e): BDHB + AZO, LAG for 30 min. Add TMP and continue LAG for 30 min. Add PYR and continue LAG for 30 min Match between M1(g), M1(h), M1(i) and M1(j)
M1(f): BDHB + AZO, LAG for 30 min. Add PYR and continue LAG for 30 min. Add TMP and continue LAG for 30 min Mismatch with calculated PXRD pattern and manual grinding in mortar and pestle
M1(g): TMP + PYR, LAG for 30 min. Add BDHB and continue LAG for 30 min. Add AZO and continue LAG for 30 min Fig. 13
and Match of M1(g), M1(h), M1(i) and M1(j) with the calculated PXRDa pattern using ball-mill grinding
M1(g): TMP + PYR, BMG for 30 min. Add BDHB and continue BMG for 30 min. Add AZO and continue BMG for 30 min 1H NMR spectrum of the quaternary cocrystal is shown in Fig. S18†
M1(h): TMP + PYR, LAG for 30 min. Add AZO and continue LAG for 30 min. Add BDHB and continue LAG for 30 min
and
M1(h): TMP + PYR, BMG for 30 min. Add AZO and continue BMG for 30 min. Add BDHB and continue BMG for 30 min
M1(i): TMP + AZO, LAG for 30 min. Add BDHB and continue LAG for 30 min. Add PYR and continue LAG for 30 min
and
M1(i): TMP + AZO, BMG for 30 min. Add BDHB and continue BMG for 30 min. Add PYR and continue BMG for 30 min
M1(j): TMP + AZO, LAG for 30 min. Add PYR and continue LAG for 30 min. Add BDHB and continue LAG for 30 min
and
M1(j): TMP + AZO, BMG for 30 min. Add PYR and continue BMG for 30 min. Add BDHB and continue BMG for 30 min
M1(k): PYR + AZO, LAG for 30 min. Add BDHB and continue LAG for 30 min. Add TMP and continue LAG for 30 min
M1(l): PYR + AZO, LAG for 30 min. Add TMP and continue LAG for 30 min. Add BDHB and continue LAG for 30 min
M2: Direct addition of BDHB + TMP + PYR + AZO, LAG for 30 min
MeOH for LAG and BMG; milling frequency is 25 Hz
2 BDHB–TMP–DPE-II–PYR (2[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) M1(a): BDHB + TMP, LAG for 30 min. Add PYR and continue LAG for 30 min. Add DPE-II and continue LAG for 30 min Fig. 14
M1(b): BDHB + TMP, LAG for 30 min. Add DPE-II and continue LAG for 30 min. Add PYR and continue LAG for 30 min Match between M2, M1(a), M1(b), M1(c), M1(d), M1(e), M1(f), M1(k) and M1(l)
and Match with calculated PXRD patterna and manual grinding in mortar and pestle
M1(b): BDHB + TMP, BMG for 30 min. Add DPE-II and continue BMG for 30 min. Add PYR and continue BMG for 30 min Fig. 15
M1(c): BDHB + PYR, LAG for 30 min. Add TMP and continue LAG for 30 min. Add DPE-II and continue LAG for 30 min Match between M1(g), M1(h), M1(i) and M1(j)
M1(d): BDHB + PYR, LAG for 30 min. Add DPE-II and continue LAG for 30 min. Add TMP and continue LAG for 30 min Mismatch with calculated PXRD pattern and manual grinding in mortar and pestle
M1(e): BDHB + DPE-II, LAG for 30 min. Add TMP and continue LAG for 30 min. Add PYR and continue LAG for 30 min Fig. 16
and Match of M1(g), M1(h), M1(i) and M1(j) with the calculated PXRDa pattern using ball-mill grinding
M1(e): BDHB + DPE-II, BMG for 30 min. Add TMP and continue BMG for 30 min. Add PYR and continue BMG for 30 min 1H NMR spectrum of the quaternary cocrystal is shown in Fig. S19†
M1(f): BDHB + DPE-II, LAG for 30 min. Add PYR and continue LAG for 30 min. Add TMP and continue LAG for 30 min
M1(g): TMP + PYR, LAG for 30 min. Add BDHB and continue LAG for 30 min. Add DPE-II and continue LAG for 30 min
and
M1(g): TMP + PYR, BMG for 30 min. Add BDHB and continue BMG for 30 min. Add DPE-II and continue BMG for 30 min
M1(h): TMP + PYR, LAG for 30 min. Add DPE-II and continue LAG for 30 min. Add BDHB and continue LAG for 30 min
and
M1(h): TMP + PYR, BMG for 30 min. Add DPE-II and continue BMG for 30 min. Add BDHB and continue BMG for 30 min
M1(i): TMP + DPE-II, LAG for 30 min. Add BDHB and continue LAG for 30 min. Add PYR and continue LAG for 30 min
and
M1(i): TMP + DPE-II, BMG for 30 min. Add BDHB and continue BMG for 30 min. Add PYR and continue BMG for 30 min
M1(j): TMP + DPE-II, LAG for 30 min. Add PYR and continue LAG for 30 min. Add BDHB and continue LAG for 30 min
and
M1(j): TMP + DPE-II, BMG for 30 min. Add PYR and continue BMG for 30 min. Add BDHB and continue BMG for 30 min
M1(k): PYR + DPE-II, LAG for 30 min. Add BDHB and continue LAG for 30 min. Add TMP and continue LAG for 30 min
M1(l): PYR + DPE-II, LAG for 30 min. Add TMP and continue LAG for 30 min. Add BDHB and continue LAG for 30 min
M2: direct addition of BDHB + TMP + PYR + DPE-II, LAG for 30 min
MeOH for LAG and BMG; milling frequency is 25 Hz

In the ternary salt–cocrystal hydrate ACF–PIP–GA–H2O (Table 3, entry 1), the three experimental PXRD patterns of M1(a), M1(b) and M2 methods (Fig. 1) overlay with each other and the calculated XRD profile from the X-ray crystal structure.1 However, there are two small peaks between 2θ = 10–12° in the case of the M1(a) method which appear to be a contamination, since these peaks are absent in the calculated PXRD line profile also. A possible reason could be that mixing of ACF + PIP and LAG grinding will lead to a salt (due to the large pKa difference between ACF and PIP), and the binary salt may not react completely to give the ternary salt–cocrystal hydrate. Indeed, the ACF-PIP binary complex was observed as a monohydrate with a strong peak at 2θ = 11.3°, which is present in the ternary product as an impurity peak. Secondly, the calculated PXRD lines are slightly shifted to larger 2θ by ∼0.5° due to the shorter d-spacing in the low temperature X-ray crystal structure of ACF–PIP–GA–H2O recorded at 100 K, while the experimental PXRD plots are at RT (300 K).


image file: d4ce00551a-f1.tif
Fig. 1 Overlay PXRD plots of ACF–PIP–GA–H2O (1[thin space (1/6-em)]:[thin space (1/6-em)]0.5[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]2) using different manual grinding conditions. M1(a): ACF + PIP, then GA (blue); M1(b) ACF + GA, then PIP (green); M2: ACF + PIP + GA (black); calculated from the X-ray crystal structure (red).

In the case of ACF–INA–GA (1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) form 1 there is excellent match between the stepwise and the direct addition grinding methods using MeCN solvent drops (entry 2, Table 3). Both methods are comparable and the experimental PXRD lines match with those of the calculated X-ray structure (Fig. 2). Form 2 of ACF–INA–GA (1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) could not be reproduced because it was formed in our reported study1 by grinding the components with 4-hydroxybenzamide additive followed by recrystallization from acetonitrile. In our grinding experiment, the result was form 1 of ACF–INA–GA (1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) by PXRD analysis. The next system described is that of a ternary cocrystal of acetazolamide, ACZ-NAM-2HP (entry 3a, Table 3). LAG grinding showed an overlay of the experimental PXRD patterns M1(a), M1(b), and M2, but they are different from that of the calculated line profile from the X-ray crystal structure (entry 1, Table S1 and Fig. S1†). PXRD overlay of additional systems discussed in the paper are shown in Fig. S2–S11 (ESI†). The presence of the three molecular components in the ternary cocrystal solid in 1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1 ratio was confirmed by 1H NMR (Fig. S12a, ESI†). A check with the binary cocrystals of ACZ with NAM and 2HP reported in the original paper36 did not indicate a match with the experimental ternary product. Higher mechanochemical pressure under ball-mill grinding furnished the ternary product which matched with that of the single-crystal product ACZ–NAM–2HP (entry 3b, Table 3 and Fig. 3). This suggests that higher grinding pressure promoted transformation to the ternary product matching with the reported single-crystal structure product. The identity of the ternary crystalline products by NMR (Fig. S12b, ESI†) suggests that under manual LAG conditions the product is a polymorph of ACZ–NAM–2HP. No further efforts were made to fully characterize this crystalline product.


image file: d4ce00551a-f2.tif
Fig. 2 Overlay PXRD plots of ACF–INA–GA (1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) form 1 using different manual grinding conditions. M1(a): ACF + INA, then GA (blue); M1(b) ACF + GA, then INA (green); M2: ACF + INA + GA (black); calculated from the X-ray crystal structure (red).

image file: d4ce00551a-f3.tif
Fig. 3 Overlay PXRD plots of ACZ–NAM–2HP (1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) using ball-mill grinding conditions. M1(a): ACZ + NAM, then 2HP (blue); M1(b) ACZ + 2HP, then NAM (green); M2: ACZ + NAM + 2HP (black); calculated from the X-ray crystal structure (red). 1H NMR of the ternary product shows the 1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1 stoichiometry in Fig. S12(b).†

The ternary cocrystal of p-sulfonamide benzoic acid (SMBA) with nicotinamide and 2-hydroxypyridone was reproduced with identical results using both LAG methods 1 and 2 (entry 4, Table 3 and Fig. 4). Whereas it is possible, and likely, that seeds or nuclei of binary and ternary cocrystals are still floating in the laboratory space for the acefylline system since these experiments were carried out recently and by the same student as the co-author of this paper, it is unlikely for crystal seeds to be still contaminating our laboratory environment for the SMBA system, since these experiments were carried out almost a decade ago by a different student but in the same laboratory room. The same can be said for the ACZ system. The success of both M1 and M2 grinding methods with SMBA ternary cocrystals suggests that a compact and expeditious protocol of direct grinding of ternary cocrystals using LAG is workable. However, not all the results were equally successful as described in the original report.37 Picolinamide (PAM) and nicotinamide (NAM) gave reproducible results in SMBA ternary cocrystals with methyl hydroxypyridone (MeHP) (entries 2 and 3, Table S1, and Fig. S2 and S3, ESI†), but isonicotinamide (INA) showed an unexpected outcome. The ternary cocrystal of SMBA–INA–2HP showed an identical microcrystalline powder using LAG methods 1 and 2 by PXRD, but these line patterns are different from that of the single-crystal ternary product (entry 5, Table 3 and Fig. 5a). The experimental PXRD pattern matches with that of the binary product SMBA–INA (1[thin space (1/6-em)]:[thin space (1/6-em)]1)37 (Fig. 5b) after manual LAG. Thus, manual LAG afforded a binary intermediate product in attempts to reproduce the ternary cocrystal. Surprisingly, ball-mill grinding of the three components showed a PXRD line pattern different from that of manual LAG grinding and the calculated PXRD pattern of the ternary cocrystal (Fig. 5c).


image file: d4ce00551a-f4.tif
Fig. 4 Overlay PXRD plots of SMBA–NAM–2HP (1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) using different manual grinding conditions. M1(a): SMBA + NAM, then 2HP (blue); M1(b) SMBA + 2HP, then NAM (green); M2: SMBA + NAM + 2HP (black); calculated from the X-ray crystal structure (red).

image file: d4ce00551a-f5.tif
Fig. 5 (a) Overlay PXRD plots of SMBA–INA–2HP (1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) using different manual grinding conditions. M1(a): SMBA + INA, then 2HP (blue); M1(b) SMBA + 2HP, then INA (green); M2: SMBA + INA + 2HP (black); calculated from the X-ray crystal structure (red). 1H NMR of the ternary product shows the 1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1 stoichiometry in Fig. S13(a) and (b)† (b) M2 LAG product matches with that of binary SMBA–INA (1[thin space (1/6-em)]:[thin space (1/6-em)]1) as confirmed by PXRD match. (c) SMBA–INA–2HP–M2 synthesized using ball-mill grinding (light green); SMBA–INA–2HP–M2 synthesized using manual grinding (black); and calculated from the X-ray crystal structure (red). 1H NMR of the ternary product shows the 1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1 stoichiometry in Fig. S13(c) and (d).†

Ternary cocrystals of bumetanide (entries 6 and 7, Table 3) were reproduced by manual LAG with identical results using M1 and M2 conditions (Fig. 6 and 7) compared to the reported study.38 However, since the calculated XRD pattern from the crystal structure is not available for these two ternary systems, the three-component 1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1 stoichiometry was confirmed by 1H NMR (Fig. S14 and S15, ESI†). Similarly BUM–INA–PCA and BUM–INA–VLA (entries 4 and 5, Table S1, Fig. S4 and S5 for PXRD, and Fig. S16 and S17 for NMR, ESI†) were reproduced in a routine manual LAG grinding procedure by both methods. Three ternary systems were reproduced with excellent overlay between the experimental and calculated PXRD patterns reported by other groups (entries 8–10, Table 3).39,40 Their PXRD overlay plots are shown in Fig. 8–10 and S6 and S7 (ESI†). With an excellent match between the calculated and experimental PXRD patterns using both M1 and M2 methods, the manual LAG grinding method was found to be reproducible (for experiments by the same student) and repeatable (for experiments done by other students and now carried out by the co-author) for a reasonable number of diverse ternary systems.


image file: d4ce00551a-f6.tif
Fig. 6 Overlay PXRD plots of BUM–2HP–PCA (1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) using different manual grinding conditions. M1(a): BUM + 2HP, then PCA (blue); M1(b): BUM + PCA, then 2HP (green); M2: BUM + 2HP + PCA (black). The 1H NMR spectrum of the ternary product is shown in Fig. S14.†

image file: d4ce00551a-f7.tif
Fig. 7 Overlay PXRD plots of BUM–INA–PASA (1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) using different manual grinding conditions. M1(a): BUM + INA, then PASA (blue); M1(b) BUM + PASA, then INA (green); M2: BUM + INA + PASA (black). The 1H NMR spectrum of the ternary product is shown in Fig. S15.†

image file: d4ce00551a-f8.tif
Fig. 8 Overlay PXRD plots of 4NBA–FA–DIB (2[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) using different manual grinding conditions. M1(a): 4NBA + FA, then DIB (blue); M1(b): 4NBA + DIB, then FA (green); M1(c): FA + DIB, then 4NBA (magenta); M2: 4NBA + FA + DIB (black); calculated from the X-ray crystal structure (red).

image file: d4ce00551a-f9.tif
Fig. 9 Overlay PXRD plots of INH–NA–SA (1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) using different manual grinding conditions. M1(a): INH + NA, then SA (blue); M1(b): INH + SA, then NA (green); M2: INH + NA + SA (black); calculated from the X-ray crystal structure (red).

image file: d4ce00551a-f10.tif
Fig. 10 Overlay PXRD plots of INH–NA–FA (1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) using different manual grinding conditions. M1(a): INH + NA, then FA (blue); M1(b): INH + FA, then NA (green); M2: INH + NA + FA (black); calculated from the X-ray crystal structure (red).

This study was extended to two quaternary systems (entries 1 and 2, Table 4), BDHB–TMP–AZO–PYR (2[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) and BDHB–TMP–DPE-II–PYR (2[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1).41 For both these systems, a few of the M1 manual LAG grinding methods produce the solid crystalline material having experimental PXRD lines matching with those of the calculated PXRD pattern (Fig. 11), while in other cases the experimental overlay was good but did not overlay on the calculated XRD lines (Fig. 12). However, upon repeating the same methods under ball-mill grinding conditions, successful formation of the quaternary cocrystal was observed whose experimental PXRD lines matched with those of the calculated crystal structure reference product (see Fig. 13 for procedures M1 g, h, i and j; except for a small systematic shift, the pattern overlays quite nicely). Similar observations on improvement in the results with BMG was noted for the second quaternary system (entry 2, Table 4, and Fig. 14–16). Representative 1H NMR spectra of the quaternary cocrystals are shown in Fig. S18 and S19 (ESI†). The differences between the MG and BMG results can be explained in terms of a more efficient mechanochemical force from the ball-mill which drives supramolecular reactivity between the four components. Any binary or ternary intermediates are converted to the final quaternary product under the strenuous BMG conditions. NMR spectra of some of the ternary and quaternary products to confirm the chemical compositions are shown in Fig. S12–S19 (ESI†).


image file: d4ce00551a-f11.tif
Fig. 11 Overlay PXRD plots of BDHB–TMP–AZO–PYR (2[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) using different manual grinding conditions. M1(a): BDHB + TMP, then PYR, then AZO (blue); M1(b): BDHB + TMP, then AZO, then PYR (green); M1(c) BDHB + PYR, then TMP, then AZO (magenta); M1(d): BDHB + PYR, then AZO, then TMP (wine red); M1(e): BDHB + AZO, then TMP, then PYR (orange); M1(f): BDHB + AZO, then PYR, then TMP (royal blue); M1(k): PYR + AZO, then BDHB, then TMP (pink); M1(l): PYR + AZO, then TMP, then BDHB (purple); M2: BDHB + TMP + PYR + AZO (black); calculated from the X-ray crystal structure (red). The calculated PXRD lines overlay with the experimental PXRD using different conditions as described in Table 4.

image file: d4ce00551a-f12.tif
Fig. 12 Overlay PXRD plots of BDHB–TMP–AZO–PYR (2[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) using different manual grinding conditions. M1(g): TMP + PYR, then BDHB, then AZO (dark yellow); M1(h): TMP + PYR, then AZO, then BDHB (grey); M1(i) TMP + AZO, then BDHB, then PYR (dark green); M1(j): TMP + AZO, then PYR, then BDHB (brown); M2: BDHB + TMP + PYR + AZO (black); calculated from the X-ray crystal structure (red). The experimental PXRD peaks using different M1 conditions as described in Table 4 match with each other in the overlay, but they are different from the M2 condition method, which matches with the calculated PXRD lines.

image file: d4ce00551a-f13.tif
Fig. 13 Overlay PXRD plots of BDHB–TMP–AZO–PYR (2[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) using ball-mill grinding conditions. M1(g): TMP + PYR, then BDHB, then AZO (dark yellow); M1(h): TMP + PYR, then AZO, then BDHB (yellow); M1(i) TMP + AZO, then BDHB, then PYR (dark green); M1(j): TMP + AZO, then PYR, then BDHB (brown); M2: BDHB + TMP + PYR + AZO (black); calculated from the X-ray crystal structure (red). The experimental PXRD peaks using different M1 conditions as described in Table 4 match with each other in the overlay and also with the M2 condition method and calculated PXRD lines. The 1H NMR spectrum of the ternary product is shown in Fig. S18.†

image file: d4ce00551a-f14.tif
Fig. 14 Overlay PXRD plots of BDHB–TMP–DPE-II–PYR (2[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) using different manual grinding conditions. M1(a): BDHB + TMP, then PYR, then DPE-II (blue); M1(c) BDHB + PYR, then TMP, then DPE-II (magenta); M1(d): BDHB + PYR, then DPE-II, then TMP (wine red); M1(f): BDHB + DPE-II, then PYR, then TMP (royal blue); M1(k): PYR + DPE-II, then BDHB, then TMP (pink); M1(l): PYR + DPE-II, then TMP, then BDHB (brown); M2: BDHB + TMP + PYR + DPE-II (black); calculated from the X-ray crystal structure (red). The calculated PXRD lines overlay with the experimental PXRD using different conditions as described in Table 4.

image file: d4ce00551a-f15.tif
Fig. 15 Overlay PXRD plots of BDHB–TMP–DPE-II–PYR (2[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) using different manual grinding conditions. M1(b): BDHB + TMP, then DPE-II, then PYR (green); M1(e): BDHB + DPE-II, then TMP, then PYR (orange); M1(g): TMP + PYR, then BDHB, then DPE-II (dark yellow); M1(h): TMP + PYR, then DPE-II, then BDHB (grey); M1(i) TMP + DPE-II, then BDHB, then PYR (dark green); M1(j): TMP + DPE-II, then PYR, then BDHB (brown); M2: BDHB + TMP + PYR + DPE-II (black); calculated from the X-ray crystal structure (red). The experimental PXRD peaks using different M1 conditions as described in Table 4 match with each other in the overlay, but they are different from the M2 condition method, which matches with the calculated PXRD lines.

image file: d4ce00551a-f16.tif
Fig. 16 Overlay PXRD plots of BDHB–TMP–DPE-II–PYR (2[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1) using ball-mill grinding conditions. M1(b): BDHB + TMP, then DPE-II, then PYR (green); M1(e): BDHB + DPE-II, then TMP, then PYR (orange); M1(g): TMP + PYR, then BDHB, then DPE-II (dark yellow); M1(h): TMP + PYR, then DPE-II, then BDHB (grey); M1(i) TMP + DPE-II, then BDHB, then PYR (dark green); M1(j): TMP + DPE-II, then PYR, then BDHB (brown); M2: BDHB + TMP + PYR + DPE-II (black); calculated from the X-ray crystal structure (red). The experimental PXRD peaks using different M1 conditions as described in Table 4 match with each other in the overlay and also with the M2 condition method and calculated PXRD lines. The 1H NMR spectrum of the ternary product is shown in Fig. S19.†

In order to define the limits of the grinding method, an unsuccessful result as a part of our study is reported in which we were unable to crystallize the expected ternary cocrystal/salt after manual LAG and BMG grinding methods. The system PRZ–FLA–PRA (Table S1, entries 8–11, ESI†) was selected to demonstrate if a ternary drug–drug salt cocrystal of piperazine ferulate with pyrazinamide42 can be crystallized using the solvent-less grinding method. In the original procedure, manual LAG using acetone in a mortar and pestle afforded the crystalline material after recrystallization from EtOH–MeCN (6[thin space (1/6-em)]:[thin space (1/6-em)]1) solvent.

Manual grinding in a mortar and pestle was performed for the ternary cocrystal salt PRZ–FLA–PRA. A ratio of 0.5[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]2 for PRZ, FLA, and PRA was ground as per the original paper.42 PXRD patterns were identical for both M1 and M2 methods, but it was mismatched with the calculated XRD pattern for the reported 1[thin space (1/6-em)]:[thin space (1/6-em)]2[thin space (1/6-em)]:[thin space (1/6-em)]2 X-ray crystal structure.42 Grinding the components in a 1[thin space (1/6-em)]:[thin space (1/6-em)]2[thin space (1/6-em)]:[thin space (1/6-em)]2 ratio was based on the X-ray crystal structure which shows one molecule of PRZ and two molecules of PRA and FLA each. Grinding in this stoichiometry yields M2 PXRD matching with M1(a) and M1(c) but different from M1(b); none of the experimental PXRDs matched with the calculated PXRD lines from the X-ray crystal structure. The PXRD patterns were the same using manual and ball-mill grinding for the 0.5[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]2 ratio, but different from that in the 1[thin space (1/6-em)]:[thin space (1/6-em)]2[thin space (1/6-em)]:[thin space (1/6-em)]2 ratio. It is noteworthy that the experimental PXRDs did not match with the calculated XRD line pattern. PXRD overlay plots are shown in Fig. S8–S11 (ESI†). Proton NMR was carried out on the crystalline powder to confirm the composition of the product in each case (Fig. S20–S23, ESI†). Attempts to crystallize the experimental product in our lab using the same solvent as mentioned in the original paper42 was unsuccessful.

Conclusion

Cocrystal synthesis via solventless grinding in a mortar and pestle or ball mill has fascinated the solid-state community for close to two decades now: initially as a synthetic method of crystalline products,2,43,44 and recently in terms of mechanistic intermediates in mechanochemistry.45 We demonstrate an excellent match between M1 and M2 grinding methods with the calculated PXRD line pattern, thus establishing the validity of the single-step multicomponent supramolecular synthesis methodology. By carrying out the multicomponent cocrystal self-assembly in a single step, the method is adaptable for one-step continuous synthesis44 of pharmaceuticals. This study validates across a wide variety of chemical systems the notion16,19 that crystal seeds or crystal nuclei are important in the formation of multicomponent crystalline products. The method was successfully upgraded from ternary to quaternary systems, thus opening its validity for a general single-step synthesis of pharmaceutical cocrystals and salts, which is now an important topic in modern solid-state pharmaceutical science. Manual LAG grinding is the first choice followed by BMG, but the severity of grinding must be controlled because excessive compression to achieve complete cocrystallization can lead to polymorphism at high pressure or result in coamorphization of the product. The ability of pharmaceutical cocrystals to tune the desired physicochemical, mechanical, and pharmacokinetic properties has already commenced the translational phase of crystal engineering from the laboratory to the marketplace. The simplicity, ease of operation, quantitative yields, high crystallinity, seamless scalability, and solvent-free synthesis in the mechanochemical route stand out as advantages of the present methodology.

Experimental section

Materials

ACF, ACZ, BUM, SMBA and INH drugs were obtained from TCI Chemicals Pvt. Ltd. (India), Yarrow Chemicals (Mumbai, India), Shanghai Xunxin Chemical Co. (China), BLD Pharmatech (India) Pvt. Ltd. and Sigma-Aldrich, respectively. Other chemicals were procured from BLD Pharmatech (India) Pvt. Ltd., Avra Synthesis Pvt. Ltd. (Hyderabad, India), Sisco Chem Pvt. Ltd. and Sigma Aldrich. Solvents were purchased from either Finar Limited or Avra Synthesis Pvt. Ltd. All chemicals/solvents were of high purity (analytical/chromatographic grade) and used as such without any further purification.

Synthesis of ternary cocrystals

Ternary cocrystal synthesis by manual grinding. The ratio of the molecular components and solvent used for cocrystal formation is kept the same as in the original paper (given in Tables 3 and S1†) for all manual grinding conditions. One prototype procedure for ternary cocrystal synthesis is elaborated here. Except for the ratio of components and solvent, the experimental procedure is similar for other systems.

ACZ–NAM–2HP: ACZ and the coformers NAM and 2HP were taken in a 1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1 ratio and EtOAc was the solvent for liquid-assisted grinding (LAG) cocrystal formation.

M2 method: ACZ was taken with both the coformers NAM and 2HP along with a few drops of EtOAc and was ground manually for 30 min. The solvent was added at periodic intervals whenever the solid mixture was appearing to be dried out due to rapid solvent evaporation in the open mortar and pestle jar.

M1(a) method: ACZ was taken with the coformer NAM and a few drops of EtOAc were added, and the solid–liquid mixture was ground manually for 30 min. Then 2HP was added with a few drops of EtOAc and again manual grinding was continued for 30 min. During both the steps, solvent was added at proper intervals whenever the solid mixture was appearing to be dried out due to rapid solvent evaporation in the open mortar and pestle jar.

M1(b) method: ACZ was taken with the coformer 2HP and a few drops of EtOAc were added, and the solid–liquid mixture was ground manually for 30 min. Then, NAM was added with a few drops of EtOAc and again manual grinding was continued for 30 min. During both the steps, solvent was added at proper intervals whenever the solid mixture was appearing to be dried out due to rapid solvent evaporation from the open mortar and pestle jar.

Ternary cocrystal synthesis by ball-mill grinding. The ratio of the molecular components and solvent used for cocrystal formation is kept the same as in the original paper (given in Tables 3 and S1†) for all ball-mill grinding conditions. One prototype procedure for ternary cocrystal synthesis is elaborated here. Except for the ratio of components and solvent, the experimental procedure is similar for other systems.

ACZ–NAM–2HP: ACZ and the coformers NAM and 2HP were taken in a 1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1 ratio and EtOAc was added as the solvent for ball-mill grinding (BMG) to facilitate cocrystal formation.

M2 method: ball-mill grinding was performed in a Retsch MM400 ball mil, equipped with a stainless steel 25 mL grinding jar and three 5 mm stainless steel grinding balls. ACZ was taken with both the coformers NAM and 2HP in the jar, 2–3 drops of EtOAc were added, and BMG was performed at a rate of 25 Hz for 30 min.

M1(a) method: ball-mill grinding was performed in a Retsch MM400 ball mill equipped with a stainless steel 25 mL grinding jar and three 5 mm stainless steel grinding balls. First ACZ was taken with the coformer NAM in the jar along with 2–3 drops of EtOAc added for LAG and grinding was performed at a rate of 25 Hz for 30 min. Then 2HP and 2–3 drops of EtOAc were added in the same jar and grinding was continued for 30 min at a rate of 25 Hz.

M1(b) method: ball-mill grinding was conducted using a Retsch MM400 ball mill equipped with a stainless steel 25 mL grinding jar and three 5 mm stainless steel grinding balls. First ACZ was taken with the coformer 2HP in the jar along with 2–3 drops of EtOAc added for LAG and grinding was performed at a rate of 25 Hz for 30 min. Then NAM and 2–3 drops of EtOAc were added in the same jar for LAG and grinding was continued for 30 min at a rate of 25 Hz.

Powder X-ray diffraction (PXRD)

Powder X-ray diffraction data were recorded on a Bruker D8 Advance diffractometer (Bruker-AXS, Karlsruhe, Germany) using Cu-Kα X-radiation (λ = 1.5406 Å) at 40 kV and 30 mA power. X-ray diffraction patterns were collected over the 2θ range of 3–50° at a scan rate of 5.34° min−1. Overlay plots of diffraction line patterns of materials synthesized using M1 and M2 methods along with the calculated line from the X-ray crystal structure were made using DIFFRACTION.EVA software.

Solution NMR spectroscopy

Solution 1H NMR spectra of crystalline materials were recorded on a Bruker Advance 400 and 500 MHz spectrometer (Bruker BioSpin, Karlsruhe, Germany) in DMSO-d6 solvent (solvent peak at 2.50 ppm). The stoichiometry of the chemical system was established by proton NMR integration.

In the other ternary systems grinding was performed in a similar manner (see Table 3 and S1† for stoichiometry and solvent used for LAG).

Quaternary systems (Table 4) are detailed in the ESI.†

Data availability

The data supporting this article have been included as part of the ESI.†

Author contributions

All authors contributed equally toward the design, experiments, and writing of this manuscript.

Conflicts of interest

The authors declare no competing financial interest.

Acknowledgements

A. K. N. acknowledges the financial and infrastructure support from UGC (through UPE, CAS, and NRC programs), DST (through PURSE and FIST programs), SERB for JC Bose Fellowship (SR/S2/JCB-06/2009), and the Institute of Eminence status for UoH (Ministry of Education). S. N. acknowledges the UGC Savitribai Jyotirao Phule PhD fellowship for single girl child (82-7/2022(SA-III)). We also thank the Institute of Eminence project (UoH-IoE-RC4-21-007) on continuous crystallization.

References

  1. S. Allu, A. Garai, V. V. Chernyshev and A. K. Nangia, Cryst. Growth Des., 2022, 22, 4165–4181,  DOI:10.1021/acs.cgd.2c00184.
  2. A. V. Trask, D. A. Haynes, W. D. S. Motherwell and W. Jones, Chem. Commun., 2006, 51–53,  10.1039/b512626f.
  3. J. Bernstein, Polymorphism in Molecular Crystals, Clarendon, Oxford, 2002 Search PubMed.
  4. G. Bolla, B. Sarma and A. K. Nangia, Chem. Rev., 2022, 122, 11514–11603,  DOI:10.1021/acs.chemrev.1c00987.
  5. S. Tothadi, A. Mukherjee and G. R. Desiraju, Chem. Commun., 2011, 47, 12080–12082,  10.1039/C1CC14567C.
  6. M. Rajkumar and G. R. Desiraju, IUCrJ, 2021, 8, 178–185,  DOI:10.1107/S2052252520016589.
  7. N. A. Mir, R. Dubey and G. R. Desiraju, Acc. Chem. Res., 2019, 52, 2210–2220,  DOI:10.1021/acs.accounts.9b00211.
  8. R. Dubey, N. A. Mir and G. R. Desiraju, IUCrJ, 2016, 3, 102–107,  DOI:10.1107/S2052252515023957.
  9. M. Böning, B. Stuhlmann, G. Engler, M. Busker, T. Häber, A. Tekin, G. Jansen and K. Kleinermanns, ChemPhysChem, 2013, 14, 837–846,  DOI:10.1002/cphc.201200701.
  10. A. Mukherjee, K. Dixit, S. P. Sarma and G. R. Desiraju, IUCrJ, 2014, 1, 228–239,  DOI:10.1107/S2052252514012081.
  11. A. S. Tatton, T. N. Pham, F. G. Vogt, D. Iuga, A. J. Edwards and S. P. Brown, Mol. Pharmaceutics, 2013, 10, 999–1007,  DOI:10.1021/mp300423r.
  12. H. Kulla, A. A. L. Michalchuk and F. Emmerling, Chem. Commun., 2019, 55, 9793–9796,  10.1039/C9CC03034D.
  13. J.-L. Do and T. Friščić, ACS Cent. Sci., 2017, 3, 13–19,  DOI:10.1021/acscentsci.6b00277.
  14. D. Braga, L. Maini and F. Grepioni, Chem. Soc. Rev., 2013, 42, 7638–7648,  10.1039/C3CS60014A.
  15. T. Friščić, in Co-crystals: Preparation, Characterization and Applications, ed. C. B. Aakeröy and A. S. Sinha, The Royal Society of Chemistry, Cambridge, 2018, ch. 4, pp. 147–193,  10.1039/9781788012874-00147.
  16. D.-K. Bučar, R. W. Lancaster and J. Bernstein, Angew. Chem., Int. Ed., 2015, 54, 6972–6993,  DOI:10.1002/anie.201410356.
  17. H. McTague and Å. C. Rasmuson, Cryst. Growth Des., 2021, 21, 3967–3980,  DOI:10.1021/acs.cgd.1c00296.
  18. K. Lin, Y. Wang and Q. Yu, J. Cryst. Growth, 2021, 570, 126229,  DOI:10.1016/j.jcrysgro.2021.126229.
  19. D.-K. Bučar, G. M. Day, I. Halasz, G. G. Z. Zhang, J. R. G. Sander, D. G. Reid, L. R. MacGillivray, M. J. Duer and W. Jones, Chem. Sci., 2013, 4, 4417–4425,  10.1039/C3SC51419F.
  20. H.-L. Cao, J.-R. Zhou, F.-Y. Cai, J. Lü and R. Cao, Cryst. Growth Des., 2019, 19, 3–16,  DOI:10.1021/acs.cgd.8b01663.
  21. D. Hasa, E. Miniussi and W. Jones, Cryst. Growth Des., 2016, 16, 4582–4588,  DOI:10.1021/acs.cgd.6b00682.
  22. Y. Yuan, L. Wang, D. Li, Z. Deng and H. Zhang, Cryst. Growth Des., 2018, 18, 7244–7247,  DOI:10.1021/acs.cgd.8b01320.
  23. D. Hasa, M. Marosa, D.-K. Bučar, M. K. Corpinot, D. Amin, B. Patel and W. Jones, Cryst. Growth Des., 2020, 20, 1119–1129,  DOI:10.1021/acs.cgd.9b01431.
  24. K. Linberg, B. Röder, D. A. Sabbagh, F. Emmerling and A. A. L. Michalchuk, Faraday Discuss., 2023, 241, 178–193,  10.1039/D2FD00115B.
  25. M. K. Jafari, L. Padrela, G. M. Walker and D. M. Croker, Cryst. Growth Des., 2018, 18, 6370–6387,  DOI:10.1021/acs.cgd.8b00933.
  26. A. Delori, T. Friščić and W. Jones, CrystEngComm, 2012, 14, 2350–2362,  10.1039/c2ce06582g.
  27. A. L. Kelly, T. Gough, R. S. Dhumal, S. A. Halsey and A. Paradkar, Int. J. Pharm., 2012, 426, 15–20,  DOI:10.1016/j.ijpharm.2011.12.033.
  28. D. Hasa and W. Jones, Adv. Drug Delivery Rev., 2017, 117, 147–161,  DOI:10.1016/j.addr.2017.05.001.
  29. Y. Xiao, C. Wu, X. Hu, K. Chen, L. Qi, P. Cui, L. Zhou and Q. Yin, Cryst. Growth Des., 2023, 23, 4680–4700,  DOI:10.1021/acs.cgd.3c00183.
  30. D. Douroumis, S. A. Ross and A. Nokhodchi, Adv. Drug Delivery Rev., 2017, 117, 178–195,  DOI:10.1016/j.addr.2017.07.008.
  31. S. N. Madanayake, A. Manipura, R. Thakuria and N. M. Adassooriya, Org. Process Res. Dev., 2023, 27, 409–422,  DOI:10.1021/acs.oprd.2c00314.
  32. F. Liu, Y. Song, Y.-N. Liu, Y.-T. Li, Z.-Y. Wu and C. W. Yan, Cryst. Growth Des., 2018, 18, 1283–1286,  DOI:10.1021/acs.cgd.7b01738.
  33. A. F. Shunnar, B. Dhokale, D. P. Karothu, D. H. Bowskill, I. J. Sugden, H. H. Hernandez, P. Naumov and S. Mohamed, Chem. – Eur. J., 2020, 26, 4752–4765,  DOI:10.1002/chem.201904672.
  34. Z. X. Ng, D. Tan, W. L. Teo, F. León, X. Shi, Y. Sim, Y. Li, R. Ganguly, Y. Zhao, S. Mohamed and F. García, Angew. Chem., Int. Ed., 2021, 60, 17481–17490,  DOI:10.1002/anie.202101248.
  35. S. P. Yelgaonkar, D. Kiani, J. Baltrusaitis and L. R. MacGillivray, Chem. Commun., 2020, 56, 6708–6710,  10.1039/D0CC02671A.
  36. G. Bolla and A. Nangia, IUCrJ, 2016, 3, 152–160,  DOI:10.1107/S2052252516000543.
  37. G. Bolla and A. Nangia, Chem. Commun., 2015, 51, 15578–15581,  10.1039/c5cc06475a.
  38. S. Allu, G. Bolla, S. Tothadi and A. Nangia, Cryst. Growth Des., 2017, 17, 4225–4236,  DOI:10.1021/acs.cgd.7b00531.
  39. S. Tothadi and G. R. Desiraju, Chem. Commun., 2013, 49, 7791–7793,  10.1039/c3cc43822h.
  40. S. Aitipamula, A. B. H. Wong, P. S. Chow and R. B. H. Tan, CrystEngComm, 2013, 15, 5877–5887,  10.1039/c3ce40729b.
  41. S. Roy, R. Gaur, M. Paul, M. Rajkumar and G. R. Desiraju, Cryst. Growth Des., 2022, 22, 7578–7589,  DOI:10.1021/acs.cgd.2c01094.
  42. X.-Z. Yu, L.-Y. Wang, F. Liu, Y.-T. Li, Z.-Y. Wu and C.-W. Yan, Cryst. Growth Des., 2020, 20, 2064–2073,  DOI:10.1021/acs.cgd.9b01710.
  43. T. Friščić and W. Jones, Cryst. Growth Des., 2009, 9, 1621–1637,  DOI:10.1021/cg800764n.
  44. M. Gajda, K. P. Nartowski, J. Pluta and B. Karolewicz, Int. J. Pharm., 2019, 558, 426–440,  DOI:10.1016/j.ijpharm.2019.01.016.
  45. K. J. Ardila-Fierro and J. G. Hernández, Angew. Chem., Int. Ed., 2024, 63, e202317638,  DOI:10.1002/anie.202317638.

Footnote

† Electronic supplementary information (ESI) available: Additional ternary systems prepared using M1 and M2 LAG/BMG methods (Table S1), PXRD overlay plots (Fig. S1–S11), and 1H NMR spectra of complexes (Fig. S12–S23). See DOI: https://doi.org/10.1039/d4ce00551a
This journal is © The Royal Society of Chemistry 2024
Click here to see how this site uses Cookies. View our privacy policy here.