Establishing the generality of cocrystallization-induced spontaneous deracemization (CoISD): predictive co-former selection, transferable process models, and one-pot operation

Cocrystallization-induced spontaneous deracemization (CoISD), introduced in 2020, couples selective crystallization of the less-soluble member of a pair of diastereomeric co-crystals with solution-phase racemization of the remaining undesired enantiomer. It is the first thermodynamic deracemization route that requires neither salt formation (unlike crystallization-induced diastereomeric transformation) nor conglomerate crystallization (unlike Viedma ripening), and it was therefore presented as a general deracemization process. Six years of follow-up work resolved the original model system's optimization and equilibrium issues (yield raised from about 51% to 73% with 80% overall deracemization and mother-liquor recycling demonstrating 100% conversion feasibility) and extended cocrystallization-induced deracemization to further targets: dihydromyricetin, first stepwise and then with a MOF-encapsulated racemization catalyst that overcomes the catalyst–co-former incompatibility, and — through a bifunctional chiral agent that merges the co-former and racemization-catalyst functions in one molecule — dihydromyricetin, $(R,S)$-carprofen, $(R,S)$-indoprofen and $(R,S)$-indobufen in one pot. The generality claim itself, however, remains unproven: only a handful of targets have been deracemized, each requiring bespoke redevelopment; the published guidance for developing a new CoISD process is a procedural chart derived from the single model system rather than a predictive criterion for chiral co-former and racemization-catalyst selection; no coupled thermodynamic–kinetic process model has been shown to transfer between systems without re-optimization; and the conditions under which crystallization and racemization can be combined in a single vessel are not established for the CoISD architecture of a separate enantiopure co-former plus a separate racemization catalyst. The problem is to close this gap.

Background

A racemate is a 1:1 mixture of the two mirror-image enantiomers of a chiral molecule, and enantiomerically pure compounds are required in particular in pharmaceutical and agrochemical manufacture. Classical resolution by crystallization separates the desired enantiomer but is capped at a 50% yield, because the other half of the material stays behind in solution. Deracemization processes overcome this cap by converting the unwanted enantiomer (the distomer) into the wanted one (the eutomer) while the separation proceeds, in principle reaching 100% yield; the degree of enantiomeric purity is quantified by the enantiomeric excess, ee, defined as $ee = |n_R - n_S|/(n_R + n_S)$, the fractional surplus of one enantiomer over the other. Existing crystallization-based deracemization routes are restricted: Viedma ripening (attrition-enhanced deracemization, a kinetic grinding process) and dynamic preferential crystallization require the compound to crystallize as a conglomerate, a mechanical mixture of enantiopure crystals, which only roughly 5–10% of organic compounds do, while crystallization-induced diastereomeric transformation (CIDT), a thermodynamic process exploiting solubility differences between diastereomeric salts, requires the target to form a salt, which excludes many neutral molecules. Cocrystallization offers a way around both restrictions: a co-crystal is a crystalline solid of two different neutral molecules held together by, for example, hydrogen bonds, so a chiral co-former can convert a neutral racemate into a pair of diastereomeric co-crystals with different solubilities, without salt formation and irrespective of conglomerate behavior. Building on earlier single-system observations (e.g., racemization combined with dual-drug co-crystal resolution of an ibuprofen–levetiracetam pair in 2018), Guillot, de Meester, Huynen, Collard, Robeyns, Riant, and Leyssens introduced in 2020 the first integrated, thermodynamically driven process of this kind, cocrystallization-induced spontaneous deracemization (CoISD): a racemic target is combined with an enantiopure co-former; the less-soluble diastereomeric co-crystal selectively crystallizes, enriching the solution in the opposite enantiomer; a solution-phase racemization reaction continuously re-equilibrates the solution; and net conversion of the distomer into the eutomer proceeds until the solid phase is enantiopure, with the crystallization driving force expressible through the co-crystal solubility product $K_{sp}$. The 2020 demonstration used the fungicide precursor $(R,S)$-4,4-dimethyl-1-(4-fluorophenyl)-2-(1H-1,2,4-triazol-1-yl)pentan-3-one (BnFTP), the chiral co-former 3-phenylbutyric acid (PBA), and the strong base 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) as racemization catalyst, exploiting the keto–enol equilibrium of the α-chiral ketone. Because DBU is protonated by the acidic co-former and racemization then only occurs at high temperature, which is counterproductive for crystallization, the process was physically split into a cold crystallization vessel (10 °C) and a hot racemization vessel (90 °C) with continuous liquid transfer between them; the runs took multiple days, yields were 39–51% of enantiopure co-crystal, and the authors identified the DBU/co-former interaction, the multi-day productivity, the not-yet-understood coupled racemization–crystallization kinetics, the untested scope across other targets and co-formers, the prospect of a one-pot variant for systems not needing high temperatures, and the incomplete approach to crystallization equilibrium as open issues. A companion kinetic study (2021) measured the BnFTP racemization rate law (first order in ee and in DBU concentration, with strong solvent effects) but left the mechanism in the acid-rich regime and a comprehensive multi-variable process model open. A 2021 optimization study on the same system showed that solvent composition (toluene with a small aqueous phase), DBU loading, co-former equivalents, and stepped cooling of the crystallization cell jointly raise the outcome to about 73% yield of enantiopure co-crystal with about 80% overall deracemization, that recycling the mother liquor allows full conversion of added racemic material, and that equilibration before filtration is attainable — but it also confirmed that a one-pot arrangement of this particular system fails because of the temperature conflict. In 2024 a second target, the dihydroflavonol dihydromyricetin (a molecule with two stereocenters that crystallizes as a racemic compound, i.e., with both enantiomers occupying the same crystal lattice rather than separating into a conglomerate), was deracemized through a stepwise alternation of co-crystal resolution and solution racemization and, in a separate study of the same year, in one pot using a bifunctional chiral agent that forms enantiospecific co-crystals with the racemic target and catalyses its solution racemization by its own alkaline catalysis — a route extended to $(R,S)$-carprofen, $(R,S)$-indoprofen and $(R,S)$-indobufen and presented as addressing the incompatibility between cocrystallization and racemization encountered in cocrystallization-induced deracemization. In 2026 encapsulating DBU inside the metal–organic framework ZIF-8 (a porous crystalline zinc-imidazolate material) was shown to isolate the base catalyst from the chiral components while preserving racemization activity, enabling cocrystallization-induced deracemization of dihydromyricetin at yields above the 50% resolution ceiling with either enantiomer selectable by the handedness of the co-former. Meanwhile, adjacent deracemization advances — mechanochemical (ball-milling) deracemization, temperature-cycling methods, and demonstrations that the interplay of racemization and crystal growth rates governs amplification in conglomerate systems — address productivity and mechanism questions for other process classes, not the generality of CoISD.

Problem Statement

Establish whether cocrystallization-induced spontaneous deracemization (CoISD) is in fact the general deracemization method its founders claimed, by closing the gap between the systems demonstrated so far (BnFTP with 3-phenylbutyric acid; dihydromyricetin with N-benzyl-1-phenylethylamine and MOF-encapsulated DBU; and dihydromyricetin, $(R,S)$-carprofen, $(R,S)$-indoprofen and $(R,S)$-indobufen with a bifunctional chiral agent that combines the co-former and racemization-catalyst functions) and the claim of applicability to a multitude of racemizable compounds, whether or not they form salts and whether or not they are conglomerate formers. The objective is to make the design and outcome of a CoISD process predictable rather than empirical. In the source's own terms the open work concerns: (i) the identification, for a given racemizable target, of a suitable chiral co-former — one whose pair of diastereomeric co-crystals with the target differ in solubility enough to give selective crystallization of the less-soluble diastereomer, the selection criterion being that observed selectivity rather than a prescribed numeric solubility gap — together with a combination of crystallization and racemization conditions whose requirements are compatible, crystallization favoring low temperature for yield while racemization often requires the opposite, and a racemization agent whose activity is not suppressed by the co-former, since in the founding system DBU is protonated by the acidic co-former and the quantity of base therefore cannot be increased; (ii) a full understanding of all process parameters, that is, of the coupled racemization and co-crystallization thermodynamics and kinetics — co-crystal phase equilibria and the solubility product $K_{sp}$, the racemization rate law, crystallization kinetics, and, where used, inter-vessel mass transfer and temperature programming — such that an optimized process can be developed for a new system and its yield, enantiomeric excess and productivity predicted without full empirical re-optimization; and (iii) the conditions under which crystallization and racemization can be combined in a single vessel, which the founders expect to be fully achievable for a process that does not require high temperatures. Whether CoISD reaches racemizable compounds beyond those demonstrated so far is part of the same question: the claim to be tested is applicability to a multitude of compounds, not to a single further example. The claim is resolved by either (a) a predictive co-former and racemization-catalyst selection rule together with a coupled racemization–crystallization process model that transfer, without full empirical re-optimization, to at least two structurally distinct racemizable targets beyond those already demonstrated, at least one of which is newly deracemized, or (b) a demonstrated bound — a class of racemizable compounds that CoISD cannot deracemize. Sub-questions (i) and (ii) form the mandatory core that decides this verdict; sub-question (iii) is a separable part of the deliverable whose resolution on a given system, whether it succeeds or fails, does not by itself settle the generality claim.

References

1. Michael Guillot, Joséphine de Meester, Sarah Huynen, Laurent Collard, Koen Robeyns, Olivier Riant, Tom Leyssens, 'Cocrystallization-Induced Spontaneous Deracemization: A General Thermodynamic Approach to Deracemization', Angewandte Chemie International Edition 59, no. 28 (2020) 11303–11306. DOI: 10.1002/anie.202002464, https://doi.org/10.1002/anie.202002464 (full text via the UCLouvain ORBi green-open-access copy, https://hdl.handle.net/2078.1/229576)
2. Michael Guillot, Joséphine de Meester, Laurent Collard, Olivier Riant, Tom Leyssens, 'Co-Crystallization-Induced Spontaneous Deracemization: An Optimization Study', Organic Process Research & Development 25, no. 4 (2021) 884–891. DOI: 10.1021/acs.oprd.0c00538, https://doi.org/10.1021/acs.oprd.0c00538 (full text via the UCLouvain ORBi copy, https://hdl.handle.net/2078.1/246592)
3. Michael Guillot, Josephine de Meester, Joop H. ter Horst, Olivier Riant, Tom Leyssens, 'Fungicide Precursor Racemization Kinetics for Deracemization in Complex Systems', European Journal of Organic Chemistry, no. 3 (2021) 473–482. DOI: 10.1002/ejoc.202001402, https://doi.org/10.1002/ejoc.202001402
4. Yujiang Xie, Shubo Liu, Jie Sun, Weiwei Tang, Junbo Gong, 'Spontaneous Deracemization of the DMY Racemic Compound with Two Stereocenters through a Stepwise Process of Cocrystallization-Induced Resolution and Solution Racemization', Crystal Growth & Design 24, no. 22 (2024) 9649–9659. DOI: 10.1021/acs.cgd.4c01144, https://doi.org/10.1021/acs.cgd.4c01144
5. Xin Su, Jie Sun, Jiaqiang Liu, Yaoguo Wang, Jingkang Wang, Weiwei Tang, Junbo Gong, 'Bifunctional Chiral Agent Enables One-pot Spontaneous Deracemization of Racemic Compounds', Angewandte Chemie International Edition 63, no. 22 (2024) e202402886. DOI: 10.1002/anie.202402886, https://doi.org/10.1002/anie.202402886
6. Xin Su, Jiaqiang Liu, Runyang Zhou, Weiwei Tang, Yanfeng Dang, Junbo Gong, 'Unlocking Efficient Spontaneous Deracemization: A MOF-Mediated Heterogeneous System Overcomes Catalytic Incompatibility in Cocrystallization', Angewandte Chemie International Edition 65, no. 27 (2026) e3464047. DOI: 10.1002/anie.3464047, https://doi.org/10.1002/anie.3464047
7. Xin Su, Shubo Liu, Jiaqiang Liu, Weiwei Tang, Junbo Gong, 'Recent advances in crystallization-induced deracemization of racemic mixtures', Separation and Purification Technology 375 (2025) 133842. DOI: 10.1016/j.seppur.2025.133842, https://doi.org/10.1016/j.seppur.2025.133842
8. Sjoerd W. van Dongen, Imane Ahlal, Michel Leeman, Bernard Kaptein, Richard M. Kellogg, Iaroslav Baglai, Willem L. Noorduin, 'Chiral Amplification through the Interplay of Racemizing Conditions and Asymmetric Crystal Growth', Journal of the American Chemical Society 145, no. 1 (2023) 436–442. DOI: 10.1021/jacs.2c10584, https://doi.org/10.1021/jacs.2c10584
9. Andrej Kolarovič, Pavol Jakubec, 'State of the Art in Crystallization-Induced Diastereomer Transformations', Advanced Synthesis & Catalysis 363, no. 17 (2021) 4110–4158. DOI: 10.1002/adsc.202100473, https://doi.org/10.1002/adsc.202100473
10. Chrystal Lopes, Yohann Cartigny, Clément Brandel, Valérie Dupray, Carole Body, Oleksii Shemchuk, Tom Leyssens, 'A Greener Pathway to Enantiopurity: Mechanochemical Deracemization through Abrasive Grinding', Chemistry – A European Journal 29, no. 35 (2023) e202300585. DOI: 10.1002/chem.202300585, https://doi.org/10.1002/chem.202300585
11. Job Gieling, Guillaume Wéry, Chrystal Lopes, Joséphine de Meester, Clément Brandel, Yohann Cartigny, Tom Leyssens, Daniel M. Baier, 'Mechanochemical Deracemization: A Sustainable Approach to Enantiopurity', Chemistry – A European Journal 31, no. 15 (2025) e202404120. DOI: 10.1002/chem.202404120, https://doi.org/10.1002/chem.202404120
12. Bram Harmsen, Tom Leyssens, 'Enabling Enantiopurity: Combining Racemization and Dual-Drug Co-crystal Resolution', Crystal Growth & Design 18, no. 6 (2018) 3654–3660. DOI: 10.1021/acs.cgd.8b00438, https://doi.org/10.1021/acs.cgd.8b00438