Photoracemization‐Based Viedma Ripening of a BINOL Derivative†

Abstract Viedma ripening is a deracemization process that has been used to deracemize a range of chiral molecules. The method has two major requirements: the compound needs to crystallize as a conglomerate and it needs to be racemizable under the crystallization conditions. Although conglomerate formation can be induced in different ways, the number of racemization methods is still rather limited. To extend the scope of Viedma ripening, in the present research we applied UV‐light‐induced racemization in a Viedma ripening process, and report the successful deracemization of a BINOL derivative crystallizing as a conglomerate. Irradiation by UV light activates the target compound in combination with an organic base, required to promote the excited‐state proton transfer (ESPT), leading thereafter to racemization. This offers a new tool towards the development of Viedma ripening processes, by using a cheap and “green” catalytic source like UV light to racemize suitable chiral compounds.


Introduction
The increasing demand of single enantiomers for industrial applicationsc ontinues to stimulate the development of feasible and economically viable methodologies to reach enantiopurity. [1,2] Besides classical resolution techniques and dynamic resolution methods, Viedma ripening has been recognized as a valid deracemization process. [3][4][5][6][7] Since its discovery in 2005, severalc onditions and parameters weres creened to improve the efficiency of this technique. [1,[8][9][10] Organic compounds, such as amino acids and their derivatives, [11,12] organometallic complexes [13] and pharmaceutical intermediates, [4,5] were successfully deracemized by using Viedma ripening. Duringt he process, as lurry of the conglomerate crystalsi nas aturated solution is intensively ground to promote breakage and dissolution as well as growth of the crystals. In this way,a ne nantiopure crys-talline final state can be achieved, in up to 100 %y ield. [14,15] The technique is based upon two essential prerequisites:t he target molecule has to crystallizea saconglomerate, that is, a mechanical mixture of enantiomerically pure crystals, and suitable racemization conditions have to be applied in solution, to continuously interconverto ne enantiomer into the other.U p to now,t he Viedma ripeningp rocess has had limited applications, given the finite number of molecules whichm eet the two prerequisites. For finding conglomerates, severals trategies have been proposed, but the scope for racemizationi ss till rather limited. Racemization strategies whichh ave been applied to Viedma ripeningt ypicallyi nvolveb ase catalysis, [5,[15][16][17] but this limits the operations to acidic chiral carbonsb earing protons that can be removed. The development of new racemizationm ethods is therefore important and is the key to extendt he applicability of Viedma ripening to severalo ther substrates. Photocatalysis is ap owerful tool to promote racemizationo fp hotosensitive organic compounds. The application of UV light to molecules having sufficient absorbance in the UV region can lead to racemization by bond breakage or by excitingt he molecular orbitals, allowing for rearrangements that are forbidden in the ground state. [18] Up to now,afew successful examples of resolution through dynamic crystallization of prochiral molecules racemizing under UV light starting from completely achiral conditions have been reported. [19,20] Here, we aim to extend the application of Viedma ripening to molecules racemizing by photocatalysis, and focus the attention on chiral BINOL derivatives. This class of compounds is widely used as catalysts in organic synthesis, as templates for chiral recognition and as building blocks for polymers and other molecules. [21,22] These compounds exhibit axial chirality, namely ab ond aroundw hich rotation is hindered due to large substituents in the closestpositions, whichare held in as patial arrangement that is not superimposable to its mirror image. [23] The photoracemizationm echanism of BINOL and itsm ethyl ether was reported by Solntsev et al. in 2009. [24] They proposed that BINOL derivatives bearinga tl east af ree hydroxyl group reach the excited state upon irradiation with UV light (337 nm), and undergo deprotonation through the so-called intermolecular excited-state proton transfer (ESPT) in the presence of as uitable base. The resulting binaphtholatei st he key achiral planar intermediate, leading to racemization after reprotonation to the hydroxyl moiety occurs. Inspired by these results we tested the combination of photoinduced racemization and Viedma ripening, and we here demonstrate its successfula pplicationt op romote the deracemization of aB INOL derivative.This research may pave the way for new Viedma ripening applications to organic conglomerate-forming compounds that can undergo photochemical racemization.

Conglomerate screening
As previously mentioned, one prerequisite for Viedma ripening is that the target compound must crystallize as ac onglomerate. Following the choice to use BINOL derivatives, based on previousf indings by Solntsev et al, [24] we focused on finding a conglomerate-forming target compound. Compound 1,2 '-(benzyloxy)-[1,1'-binaphthalen]-2-ol ( Figure 1) was previously reportedt om anifest conglomerate behavior after crystallization in diethyl ether/hexane. [25] However,i na na ttempt to reproduce this experiment,n o conglomerate formation was observed by using diethyle ther/hexane as as olvent/antisolvent composition. In fact, the compound crystallizeda sanew racemic compound form, Figure 2a. Therefore, we searched for ac rystal form of 1 that does showconglomerate behavior.R ecently,o ur group has reported as ystematic approach for predicting suitable coformers to obtain stable cocrystals, some of whichp otentially being conglomerates. [26] However,g iven the complexity that such coformers might add to the system, ac onvenient ands impler route to explore conglomerate formation was to evaluate al ibrary of different sol-vents that may give rise to possible conglomerate solvates. Thus far,o nly one successful Viedma ripening deracemization involving as olvate had previously been reported, although that wasarare case of ac onglomerate co-crystal salt solvate. [27] Therefore, this work would represent an ew example in which Viedma ripening was applied to ac onglomerate solvate. To achieve this, ac rystallization screening was performed by using approximately thirty solvents. As ar esult, more than 70 %o ft he crystallization experiments led to the stable, ansolvate racemic compound form, Figure2a. For ap hase diagram study of compound 1,s ee the report by Hoquante et al. [28] However, we found that compound 1 forms ac onglomerate solvate when crystallization occurred in tolueneo rchlorobenzene,d isplaying in both cases as imilar crystal structure. As showni nF igure 2b,t wo molecules of 1 coordinate one molecule of toluene in the asymmetricunit. For all further racemization and deracemization experiments performed in this work toluene was selected as the solvent of choice.
However,i faslurry of 1 was left longer than 24 hours in toluene, completec onversion to the mosts table racemic compoundf ormo ccurred, which indicatest hat the conglomerate is in fact metastable. As imilar situation was previously reported by Spix et al.,w ho deracemized am etastable conglomerate of glutamic acid using Viedma ripening. [29] Racemization study Xenon,m ercury,d euteriuma nd LED lamps were tested as potentiall ight sourcesf or the racemization. Among them,t he xenon UV lamp provedt ob et he best solution,h aving ab road emission band covering both UV and visible regions. The UV/ Vis absorption spectrumf or 1 shows mainly three bands:a strong band at 230 nm, and two weaker bands at 285 and 335 nm, as also previously reported by Solntsev et al. [24] The solvents in which 1 crystallizes as ac onglomerate, toluene or chlorobenzene, also absorb in the UV regionu pt o2 85 nm (see UV/Vis absorption spectrum in the Supporting Information, Figure S6). Therefore, parallel experiments were performed in both quartz and glass vials, as the latter show ac utoff of radiation around3 00 nm, and the respective effects and differences in the outcome of the reactions were analyzed.R egardlesst he type of vialu sed, racemization was observed in all experiments. Recalling the mechanism reported for BINOL and  BINOL ethers, [24] we report here as imilarp athway fort he target compound 1 (Figure 3).
Irradiation by UV light promotes compound 1 to its excited state, on which the ESPT can occur.D eprotonation by suitable bases leads to aphenolate charged species, which is in equilibrium with itsa chiral planar quinoidal form. Subsequently,r eprotonation to the hydroxyl moiety occurs, resultingi nr acemization of the starting enantioenriched material. Some common organic bases,w ith pK a values between 11 and 14, were selected to be screened for their ability to efficiently perform the ESPT.F igure 4p resents an overview of racemization rates observed with different bases. Experiments were performed in toluene in stoichiometric ratio or excess with respectt ot he substrate by using ax enon lamp as the UV/Vis-light source. Pyrrolidine and racemic sec-butylamine were finally considered to be the best candidates, most likelyb ecause other bases with pKav alues above 11 are too stronga nd not suitable for allowing reprotonationtoo ccur.
Twoi ndependent experiments using only the base or UV light were performed.N oc hange in the enantiomeric excess was observed, which remained constanta t1 00 %, demonstrating that the combination of base and UV light is essential for the racemization to occur.

Deracemization using Viedmar ipening
Full conversion of the metastable conglomerate to the more stable racemic compound occurred in about 24 hours, and thus the Viedma ripening has to take place well within this time frame. Experiments starting from racemic mixtures were not successful,d ue to the too long deracemization time.
Viedmar ipening experiments were therefore performed starting from as calemic mixture (around 20 % ee)i no rder to complete the deracemization more quickly. Figure 5shows the evolution of the ee duringo ne of the Viedma ripening experiments.T he deracemization curve displays an initial steep increase of the ee,p robably because of ar ise of the temperature due to the irradiation with UV light. After the temperature stabilizes to av alue of about 40 8C, the amplification of the ee to reach (nearly) full enantiopurity is decelerated. An ee of 96 % was achieved in 4hours. Analyses of the crystalline phase by SFC (Supercritical Fluid Chromatography), coupled with UV/Vis and mass analyzer detectors, showed that no side products were present. XRPD measurements of the crystalline compound, before and after the experiments, confirmed that during the deracemization the crystals were consistentlyi n  Continued grindingu nder racemization conditions after deracemization was achieved, however,l ed to af ull conversion to the stable racemic compound, with ac onsequentd ropo f the ee and loss of enantiopurity.A lthough ideally stoppingt he racemization,t hat is, interrupting the irradiation of the compound by UV light by simply switching off the UV lamp, would be sufficient to preserve enantiopurity of the crystalsand allow for their collection, in the presentc ase some small conversion to the racemic compound still occurred. This is due to ac rystallization event taking place when the irradiation is stopped and the temperature decreases, allowing for the formation of crystalso ft he opposite enantiomer and also of the racemic compound. Therefore, an ideal experiment is stopped by 1) harvesting and washing the solid as soon as 100 % ee is reacheda nd 2) recrystallizing the final product in ad ifferent solventt oa llow for the formation of am ore stable form for the enantiomerically pure crystals.
In two cases, Viedmar ipening experiments were startedd irectly on the crude mixture from the synthesis of 1,socontaining residues such as the startingm aterial (BINOL) and the disubstituted side product.I np resence of these "impurities", the experiments proceeded to completion in am uch shorter time, leadingi nb oth cases to the final compound with (R)-configuration. Steendam et al. already showed that the presence of small, chirali mpurities can steer the outcomeo ft he deracemization and can speed up the process. [30] As recently reported, a very small amount of chiral impurities can already be sufficient to halve the deracemization time. [31]

Conclusions
We have reported here the first example of ap hotoassisted Viedmar ipening deracemization using UV light. The compoundi nvestigated in this paper proved to be am etastable conglomerate solvate. Ax enon lamp was the optimum light sourceb ecause it has ab road emission spectrum covering all UV absorption regionso ft he compound. Several bases were evaluated for their ability to promote the subsequent ESPT, necessary for the racemization to occur.U nlike at ypical Viedmar ipeninge xperiment, an initial enantiomeric excess was required to speed up the process to completion before the change of the metastable conglomerate crystal phase to the stable racemic compound form occurred. This shows the importance of detecting the differentcrystal forms of as pecific compound when dealing with possible metastable phases. Chiral impurities might be very helpful in the presence of metastable conglomerates, when deracemization has to be completed in as hort time frame.T he application of Viedma ripeningm ay be extended to other conglomerate-forming chiral organic compounds that show racemization under UV light irradiation.

Racemization experiments
Experiments were performed at room temperature in 20 mL glass vials with cutoff of radiation at 300 nm, as well as quartz vials. 25 mg of (R)-1 were dissolved in 3mLo ft oluene, after which 1equiv of the base was added to promote the "proton transfer" mechanism, and therefore the racemization. The solution was irra-diated with aU Vx enon lamp (150 W) and samples were taken every 10 minutes during the first hour,a nd subsequently every 30 minutes. The lamp was positioned at ad istance of about 20 cm from the vial used, to prevent ad rastic increase in the temperature at which the experiments took place. The temperature measured during the experiments ranged from an initial 22 8Ct oas teady value of 40 8C. The enantiomeric excess was measured by using a Waters SFC (Supercritical Fluid Chromatography), equipped with both UV and mass spectrometer detectors, on aL UX amylose-1 column with af low rate of 0.5 mL min À1 .R etention times were 2.2 min and 3.2 min for the (S)-and (R)-enantiomers, respectively. Chromatograms were recorded at aw avelength of 230 nm, corresponding to the maximum absorption peak of the BINOL derivative.

Deracemization experiments
The experiments were performed at room temperature in 20 mL glass vials with cut-off of radiation at 300 nm, as well as quartz vials. An initial enantiomeric excess (ca. 20 %) was necessary to ensure that the deracemization process would reach completion before the conversion to the more stable racemic compound crystals took place. As lurry was made from 1.2 go frac-1,0 .3 go f( R)-1 and 1.5 mL of toluene. Following ah omogenization time of about 15 minutes, 50 mLo ft he chosen base and 1gØ2mm glass beads were added. The suspension was stirred at 700 rpm and then subjected to UV-light exposure by using aU Vx enon lamp (150 W). The lamp was positioned at ad istance of about 20 cm from the vial used and the temperature measured during the experiments ranged from 22 8Ct oasteady value of 40 8C. Placing the lamp closer to the vial resulted in complete dissolution of all solid material, due to the heat produced by the lamp (nearly 50 8C). Samples were collected every 10 minutes for the first hour and then every 30 minutes, filtered off directly from the reaction slurry and analyzed after being carefully washed with one or two drops of toluene. At the end of the experiments, roughly 0.45 go fe nantiopure material was collected, which represents an et gain of half of the amount of enantiopure compound initially added. The enantiomeric excess was measured by using aW aters SFC (Supercritical Fluid Chromatography), equipped with both UV and mass spectrometer detectors, on aL UX amylose-1 column with af lowrate of 0.5 mL min À1 .R etention times were 2.2 min and 3.2 min for the (S)and (R)-enantiomers, respectively.C hromatograms were recorded at aw avelength of 230 nm, corresponding to the maximum absorption peak of the BINOL derivative. No peaks of possible side products forming during the exposure time were detected after SFC analyses. Characterization of the final enantiopure product with XRPD showed an identical diffractogram as for the starting material, indicating that the experiments were stopped in time before the crystals would undergo any conversion to the racemic compound form. XRPD diagrams of all solid samples collected during the deracemization experiments were also analyzed, to prove that the toluene solvate compound was not converted. CCDC 1940976, 1940977, 1940978, 1940979, 1940980 and 1940981 contain the supplementary crystallographic data for this paper. These data are provided free of charge by The Cambridge Crystallographic Data Centre.