(Thio)urea Containing Chiral Ammonium Salt Catalysts

(Thio)‐urea‐containing bifunctional quaternary ammonium salts emerged as powerful non‐covalently interacting organocatalysts over the course of the last decade. The most commonly employed catalysts in this field are either based on Cinchona alkaloids, α‐amino acids, or trans‐cyclohexane‐1,2‐diamine. Our group has been heavily engaged in the design and use of such catalysts, i. e. trans‐cyclohexane‐1,2‐diamine‐based ones for around 10 years now, and it is therefore the intention of this short personal account to provide an overview of the, at least in our opinion, most significant and pioneering achievements in this field by looking on catalyst design and asymmetric method development, with a special focus on our own contributions.


Introduction
The use of chiral quaternary ammonium salt ion pairing catalysts (Quatsalts) for stereoselective applications has been established as one of the fundamental catalysis principles in asymmetric organocatalysis. [1,2] Commonly classified as chiral phase-transfer catalysts (PTCs), [1][2][3] these non-covalently interacting catalysts have been successfully applied for a multitude of different organic transformations for more than 40 years, since the seminal contributions by Wynberg, [4] Dolling, [5] and O'Donnell [6] demonstrated the potential of Cinchona alkaloid-based Quatsalts I for asymmetric epoxidations and especially for enantioselective α-alkylations of prochiral starting materials. During the 1990s, and long before the hype on modern organocatalysis started, this concept already experienced its first peak with a variety of pioneering contributions using Cinchona alkaloid-derived Quatsalts. [6,7] At the end of the last millennium, Maruoka then introduced his highly active and versatile trademark binaphthyl-based ammonium salts II (the so-called Maruoka catalysts), [8] which have since then emerged as the second privileged class of chiral ammonium salt PTCs besides Cinchona alkaloids (Scheme 1A). Over the years a variety of other chiral backbones [9][10][11] have also very successfully been utilized to access powerful catalysts (e. g. Quatsalts based on tartaric acid [9] ) and the use of chiral ammonium salt catalysts has contributed to the advancement of asymmetric (organo)catalysis in a unique manner. [1,2] When discussing the use of Quatsalts as ion pairing (phase-transfer) catalysts, it also has to be stated that the elucidation of the true mechanistic scenarios and the interactions between the catalysts and the substrates is not trivial. First of all, although most reactions are commonly carried out under "classical" biphasic phase-transfer conditions, the last years also witnessed the introduction of (base-free) protocols operating under homogeneous conditions, thus making a classical phase-transfer mechanism less plausible. In addition, more and more studies underscoring a scenario where the interaction between the quat. ammonium group and the anionic/ negatively polarized group of a substrate/reagent is not just based on Coulombic attraction, but more likely a consequence of H-bonding between the N + -C α H and the substrates, have been reported. [12] Furthermore, ever since the seminal reports in this field, [4][5][6] it has been well-documented that the presence of a second catalytically competent functional group, i. e. a hydrogen-bonding motive, can have a very beneficial effect on the catalytic properties of the Quatsalts, [13] which of course makes the mechanistic understanding even more complex.
For decades the main focus in bifunctional ammonium salt catalysis was on Quatsalts containing a free OH-group as the H-bonding motive and unique transformations using OH-containing Cinchona alkaloid-based quatsalts I [4][5][6]14] as well as OH-containing Maruoka type catalysts III [15] have been introduced (Scheme 1A).
Surprisingly however, the incorporation of other Hbonding groups, i. e. the use of (thio)-ureas as powerful dual H-bonding motives, has only lately attracted the attention of scientists working in this field (Scheme 1B). [16][17][18][19][20][21][22][23][24][25][26][27][28] Very impressively however, right from the first reports describing the synthesis and use of (thio)-urea-containing chiral ammonium salts IV-VII, [16][17][18][19] the unique catalytic properties of these catalysts have been well-documented. Accordingly, it comes as no surprise that a variety of different catalysts (based on different chiral backbones) for a broad diversity of asymmetric applications have been introduced over the last decade. Our group has had a strong research focus on design and applications of such catalysts for almost 10 years now. [19,20,24,25] It is therefore the intention of this personal account to provide an overview of the, at least in our opinion, most significant and pioneering achievements in this field by looking on catalyst design (Chapter 2) and asymmetric method development (Chapter 3) with a special focus on our own contributions (i. e. with respect to applications using the catalysts developed in our group).

Catalyst Classes
As stated in the introductory chapter, Cinchona alkaloids I and axially chiral binaphthyl derivatives II and III represent the privileged sources of chirality for classical and free OHcontaining Quatsalts. While Cinchona alkaloids have also been used to access (thio)-urea containing Quatsalts IV and V (as discussed in chapter 2.

P e r s o n a l A c c o u n t T H E C H E M I C A L R E C O R D
ammonium salt hybrid catalysts. [29] Instead however, α-amino acids (α-AA; chapter 2.2.) and trans-cyclohexane-1,2-diamine (DACH; chapter 2.3.) have very successfully been used for this purpose. Interestingly, neither of these two easily available sources of chirality has been commonly employed for classical monofunctional Quatsalts before, but, as outlined in the following chapters, their use to access the powerful bifunctional ammonium salt catalysts VI and VII has contributed significantly to the advancement of the field. One point that has to be emphasized when discussing these different classes of bifunctional Quatsalts is, that the incorporation of both catalytically relevant functional groups is not always that trivial and the strategies had to be individually optimized for the given chiral backbones. In principle there are two approaches which have been both successfully applied in the past: Either installation of the Hbonding motive first, followed by a final quaternization (Strategy A, Scheme 2) or the inverse order of manipulations (B, Scheme 2).

Cinchona Alkaloids
The first report describing the synthesis and use of a thiourea-containing chiral ammonium salt dates back to 2010 when the groups of Fernandez and Lassaletta described the synthesis of one derivative of a 6'-thiourea-containing Cinchona alkaloid-based ammonium salt IV (Scheme 3). [16] In this milestone report they demonstrated the general potential of such bifunctional catalysts to facilitate reactions where analogous monofunctional derivatives fail by carrying out the asymmetric 1,4-addition of cyanides to nitroolefines (vide infra, Scheme 7A). However, it also already became obvious at that time, that the synthesis of these bifunctional catalysts is not a trivial task. Only one derivative could be Scheme 1. Privileged classes of monofunctional and free OH-containing Quatsalts (A) and the most prominent classes of (thio)-urea containing bifunctional ammonium salt catalysts (B).

Scheme 2.
The two most commonly applied assembly strategies to access (thio)-urea containing Quatsalts.

P e r s o n a l A c c o u n t T H E C H E M I C A L R E C O R D
obtained when using route A (thiourea first, quaternization second; please compare with Scheme 2) as the final quaternization was found to be difficult in the presence of the thiourea motive. Very recently, our group has readdressed the challenging syntheses of these interesting catalysts and succeeded in developing an alternative strategy following the inverse assembly approach (route B). [20] This allowed us to access a rather diverse library of 6'-urea-and thiourea containing catalysts IV, although it has to be admitted that these Quatsalts have so far not been satisfyingly used for any other applications apart from Fernandez and Lassaletta's initial one. [16] The first reports on 9-(thio)-urea-containing Cinchona alkaloid Quatsalts V appeared in 2012/13, when the groups of Dixon and Smith first investigated the synthesis and use of such bifunctional catalysts (Scheme 3). [17] Following assembly approach A, they succeeded in accessing diversely functionalized catalytically promising derivatives and over the following years such Quatsalts have been developed further and very successfully used by several groups for numerous asymmetric applications. [21]

α-Amino Acids
In general, naturally occurring α-amino acids have not emerged as a commonly used source of chirality to access classical monofunctional Quatsalts. Impressively however, in 2013 Zhao's group first reported the use of α-AA to access the powerful bifunctional Quatsalts VI (Scheme 4). [18] By developing a highly modular and robust synthesis protocol consisting of (thio)-urea installation first, followed by final quaternization, they succeeded in introducing a very diverse library of different catalysts and the analogous phosphonium salts are accessible as well. [3] Over the last years, catalysts VI have been very successfully used by several research groups for numerous impressive applications [22,23] and it is without doubt that these systems nowadays represent one of the most powerful classes of bifunctional Quatsalts that are available and their modular synthesis makes a rapid screening and finetuning for a given target reaction possible.

trans-Cyclohexane-1,2-Diamine
trans-Cyclohexane-1,2-diamine has been established as a privileged source of chirality in asymmetric (organo)-catalysis (e. g. Takemoto's trademark thiourea catalyst [30] represents one of the most commonly used organocatalysts nowadays). Surprisingly however, it has not been systematically utilized as a chiral backbone for Quatsalts until we first introduced the (thio)-urea containing ammonium salts VII in 2014 (Scheme 5). [19] Interestingly, shortly after we published our first report, Massa's group also reported a single example of such a catalyst [26] and over the last years other groups became interested in these bifunctional Quatsalts as well. [27,28] Over the years we have systematically explored the use of these catalysts for different CÀ C bond formations (often in close collaboration with Massa's group) [24] as well as asymmetric heterofunctionalizations. [25] The synthesis of a diverse library of derivatives was only possible when carrying out the (thio)urea installation after the quaternization (route B), while the opposite order of assembly, which was very successfully used for the AA-based catalysts VI (Scheme 4), was found to be rather limited for this skeleton because of side reactions of the H-bonding motive under the alkylation conditions. Upon developing a carefully optimized protecting group strategy we succeeded in introducing a robust procedure to access a broad variety of differently functionalized (thio)-urea containing Quatsalts VII. [19,24,25] Very recently we also managed to access the guanidine derivatives VII-NH [31] and the selenourea-containing analogs VII-Se as well [32][33][34] but it should be admitted that these derivatives have so far not been that exhaustively tested for their catalytic potential as compared to the established (thio)-urea Quatsalts VII.

Alternative Backbones
The three classes of bifunctional Quatsalts discussed so far clearly present the most commonly used ones. Nevertheless, over the course of the last years our group also investigated the suitability of alternative chiral skeletons with different linker lengths. As outlined in Scheme 6, other chiral backbones could be successfully utilized to access the (thio)-urea containing Quatsalts VIII-XI (in each case utilizing assembly strategy B). [19,31] Unfortunately however, in all reactions investigated so far, these catalysts were less promising than the DACH-based Quatsalts VII.

Selected Applications
The three major classes of (thio)-urea containing bifunctional Quatsalts introduced so far (chapter 2) have been successfully utilized for a broad diversity of different asymmetric target transformations over the course of the last decade. It is not the intention of this article to provide a detailed overview about all these applications, but we rather wish to show selected examples of the, at least in our opinion most important, early achievements by others, as well as we will discuss some of our own contributions which we are personally most excited about.

CÀ C Bond Formations
Asymmetric CÀ C bond forming reactions, i. e. α-alkylations of prochiral pronucleophiles and conjugate addition reactions, have for decades been amongst the most privileged applications of classical chiral Quatsalts. [1,2] Thus, it also comes as no big surprise that an asymmetric CÀ C bond formation, i. e. a conjugate addition reaction, was the reaction of interest in the seminal paper describing the first synthesis and use of a (thio)-urea containing chiral Quatsalt. [16] In 2010, the groups of Lassaletta and Fernandez investigated the vinylogous addition of TMSCN to nitroolefines 1 using different chiral organocatalysts and found that the bifunctional catalyst IV-1 clearly outperformed analogous ammonium salts missing the thiourea group as well as analogous thiourea derivatives missing the quat. ammonium group (Scheme 7A). [16] This early contribution demonstrated impressively how powerful such bifunctional catalysts can be but, as already stated before (chapter 2.1), the synthesis of catalysts IV was found to be tricky and thus, to the best of our knowledge, no further applications using these Quatsalts have been reported since then. Two years after this report, the Dixon group first described the 9-urea-containing Quatsalt V-1 and used it very successfully for nitro-Mannich Scheme 6. Alternative chiral backbones investigated by our group. Scheme 7. Selected pioneering asymmetric CÀ C bond forming applications using Cinchona alkaloid-and α-AA-based Quatsalts.

P e r s o n a l A c c o u n t T H E C H E M I C A L R E C O R D
reactions to access products 5 (Scheme 7A). [17a] Again, the importance of the bifunctional nature of these catalysts was well demonstrated and over the following years other groups used analogous nitro-Mannich reactions very successfully to develop and benchmark novel (thio)-urea containing Quatsalts as well. [18a,21a-c] Shortly after Dixon's report, the Smith group used the analogous catalyst V-2 for the enantioselective 5-endo-dig cyclization of isonitriles 6, a strategy which provides a direct entry to access substituted enantioenriched indolenines after reduction of the imine functionality of product 7.
[17b] Building on these pioneering contributions by Dixon's and Smith's groups, Quatsalts V became more and more heavily investigated over the course of the next years, resulting in the development of several highly impressive CÀ C bond forming approaches (e. g. asymmetric α-alkylations have been very successfully facilitated with these catalysts, i. e. by Connon's group [21e,g] ). α-AA based catalysts VI were first reported by Zhao's group in 2013 [18] who initially used these catalysts for nitro-Mannich reactions [18a] followed by the synthesis of chiral N,S-acetals (see section 3.2., Scheme 9A). [18b] These catalysts, as well as their analogous phosphonium salts, [3e] have been utilized for numerous applications over the course of the last decade [22,23] and one example that we wish to highlight herein is the use of catalyst VI-1 for 1,3-dipolar cycloadditions between glycine Schiff bases 9, which serve as azomethine ylide-precursors, and Michael acceptors 8 giving access to chiral spirocyclic pyrrolidines 10 with four consecutive stereogenic centers with excellent levels of stereocontrol (Scheme 7B). [23b] In 2014, the Massa group first reported the aldol-initiated cascade reaction between ortho-cyanobenzaldehyde 11 and the ketimine-containing glycine Schiff base 12 in the presence of chiral ammonium salt catalysts. 26 This reaction yields a cyclic imidate first, which can then be hydrolyzed to the phthalide 13 (Scheme 8). In their initial studies they used Dixon's catalyst V-1 which allowed for moderate levels of enantioselectivity (e.r. = 75 : 25). This report overlapped with our early investigations focusing on the design and use of the DACH-based catalysts VII and our groups thus started a long-lasting collaboration aiming on the use of catalysts VII for aldol-initiated cascade reactions of acceptors 11 with different nucleophilic species. [24] First, we were able to identify the catalyst derivative VII-1 which allowed for high enantioselectivities for the synthesis of phthalides 13. [24a] In addition, we also showed that the regioisomeric catalyst derivative VII-2 allowed for promising enantioselectivities for the addition of malonates 14 to acceptors 11. [24b] In this case, the reaction does not give the phthalide-containing product but provides an entry to the isoindolinone core (compound 15 c) instead. Formation of the latter can be rationalized by the fact that the α-acidic primary addition product 15 a undergoes a Dimroth-type rearrangement under the basic reaction conditions. Noteworthy, it is therefore not the CÀ C bond forming step where the absolute configuration of the stereogenic center is controlled, but rather the aza-Michael type cyclization of the intermediate 15 b. Primary reaction products like compound 15 c can then be further converted into potentially biologically active targets, as exemplified for the straightforward synthesis of (S)-PD 172938, a potent dopamine D4 ligand [35] (it should be emphasized that the establishment of these transformations was primarily done by Massa's group while we mainly focused on catalyst optimization).
Over the years both our teams have worked together on other conceptually similar applications as well [24] and our group has also successfully utilized catalysts VII for asymmetric Michael type reactions [24a] as well as nucleophilic aziridine ring opening reactions, [36] to mention two further examples of asymmetric CÀ C bond forming reactions where these catalysts showed their potential.

C-Heteroatom Bond Formations
Quatsalt-catalyzed C-heteroatom forming reactions, i. e. asymmetric α-heterofunctionalizations of prochiral nucleophiles, have been intensively investigated in the past. [37] Accordingly, it is not surprising that (thio)-urea-containing Quatsalts have also demonstrated their potential for conceptually different asymmetric C-heteroatom bond formations ever since the early reports in the field. In 2013 already, Zhao's group showed that their α-AA-based Quatsalts VI allow for high enantioselectivities in the asymmetric addition of thiophenols 17 to imines 16 (giving the N,S-acetals 18; Scheme 9A). [18b] More recently, Connon's group then showed that the carefully designed Cinchona alkaloid-based urea Quatsalt V-3 holds potential to control the desymmetrization of mesoanhydrides 19 with KF (Scheme 9B). [21d] Yielding the acylfluoride 20 initially, this species is then trapped with MeOH giving the halfesters 21 with moderate enantioselectivities. Although the observed levels of enantiodifferentiation still leave room for improvement, this report provides a promising proof-of-concept for the potential of bifunctional Quatsalts to control simple nucleophiles like KF in asymmetric transformations.
Our group has been interested in using our DACH-based catalysts VII for asymmetric α-heterofunctionalizations since we started investigating them. [19,25] The initial application that we tested, once we had our first library of catalysts available, was the asymmetric α-fluorination of various βketoesters 22 [38] using NFSI (23) as an electrophilic fluorinating agent (Scheme 10A). [19] Here, the urea-based catalyst VII-3 was found to be the best suited, delivering the products 24 with good levels of enantioselectivities under classical biphasic conditions. Right from the beginning, we were also interested in the activation modes of these catalysts. The beneficial effect of the bifunctional nature of Quatsalts VII was well-proven, as simplified control catalysts missing either of the two functional groups gave significantly reduced selectivities only. In detailed experimental and computational mechanistic studies together with the Vetticatt group we were able to support the proposal that catalysts VII operate via a highly ordered bifunctional activation mode. Our data suggest a scenario where H-bonding between the urea group and the enolate of the β-ketoester on the one hand, and nonclassical H-bonding between the quat. ammonium group's α-H and the Lewis basic S=O group of the NFSI on the other hand, result in a highly order unique transition state. [12c] Besides this fluorination, we also looked at the analogous α-chlorination (Scheme 10B). [25b] Although in this case the selectivities were generally lower, we made one very interesting observation hereby, as this was the only application so far, where a catalyst containing a simple trimethylammonium group (Quatsalt VII-4) was found to be more selective as

P e r s o n a l A c c o u n t T H E C H E M I C A L R E C O R D
compared to the otherwise preferred ones with a benzylic ammonium group (see derivatives VII-1-3,5).
Besides α-halogenations of β-ketoesters, we also became interested in investigating analogous α-hydroxylations. We first tested the use of oxaziridines 27 as electrophilic Otransfer agents (Scheme 11A). [25a,39] Hereby we found that the use of catalyst VII-5 allows for the synthesis of products 28 with very high enantioselectivities under base-free conditions. Furthermore, the use of 2 equiv. of racemic oxaziridines 27 was found beneficial, as the reaction proceeded with a simultaneous kinetic resolution of the O-transfer reagents, thus resulting in a very efficient overall process where the catalyst not only controls the face-selectivity of the O-transfer process to the ketoesters 22, but also shows a pronounced match/mismatch behavior with respect to the oxaziridine 27. Further mechanistic studies have been carried out and DFT analysis also showed that a similar mode of activation as for the fluorination can be proposed (Scheme 11A). [25a] At this point it should be noted that we also tested if a similar strategy may be applicable to resolve analogous racemic aziridines. [36] Unfortunately however, although some selectivity was observed, this process was overall much less satisfactory as the herein discussed oxaziridine resolution.
In a further attempt to replace oxaziridines 27 for more simple oxidants we recently realized that it is possible to carry out the asymmetric α-hydroxylation of compounds 22 with H 2 O 2 as the oxidant in the presence of aldimines 29 as catalytically competent O-transfer shuttles (Scheme 11B). [25c] Initially, we hypothesized that imine 29 undergoes in situ oxidation to oxaziridine 27, which then serves as the Otransfer agent in analogy to our previous protocol (Scheme 11A). [25a] However, control experiments showed that oxaziridines are not involved in this reaction. In contrast, we have strong evidence that the ammonium iodide first undergoes oxidation to a catalytically competent hypoiodite species, a catalytically unique concept that has recently attracted more and more attention, [40] especially since Ishihara's seminal contributions. [41] This hypoiodite then reacts with imine 29 to form a hitherto not fully elucidated active OH-transfer agent. Overall, this protocol worked well for numerous βketoesters 22, although the selectivities were slightly lower as compared to the oxaziridine approach.
One point that should be addressed is, that for all our applications so far, [19,24,25,36] the urea-based DACH Quatsalts VII were found to be more selective and often more reactive than the analogous thiourea derivatives. For the later ammonium hypoiodite-mediated reaction this was also the case. [25c] Interestingly however, when using our novel selenourea derivative VII-5-Se (please compare with Scheme 5), this catalyst was found to be even slightly more selective than our parent urea Quatsalt VII-5, [32] which thus makes the selenourea Quatsalts VII-Se promising candidates for future investigations.
Although most of the applications using DACH-based catalysts VII so far came from our group (often in collaboration with Massa's team) we were recently also very excited to learn that the groups of Tsukano, Papai, and Takemoto impressively used thiourea-containing DACH Quatsalt VII to control asymmetric O-alkylations (Scheme 12). [28] After intensive catalyst development they were able to introduce a remarkable procedure for the dynamic kinetic resolution of racemic α-chloro-lactones 31 via reaction with enol derivatives 30 in the presence of the advanced catalyst derivative VIIf. The high potential of this novel acetalization methodology, which was also supported by detailed mechanistic studies, was impressively underscored by employing it as a key step in the total synthesis of the strigolactones. [28]

P e r s o n a l A c c o u n t T H E C H E M I C A L R E C O R D
mentary alternatives compared to classical quaternary ammonium salts over the last decade. Reliable synthesis routes to access the three most commonly employed catalyst classes based on Cinchona alkaloids, α-amino acids, or transcyclohexane-1,2-diamine have been introduced and these catalysts have been very successfully utilized for a variety of different asymmetric transformations. Several research groups, including ours, have been engaged in catalyst design for more than ten years now and a variety of the catalysts introduced so far are nowadays tested and used on a regular basis by others focusing on asymmetric method development already. We are however convinced that this field is still far from being mature, as we are confident that more powerful catalyst platforms, either based on alternative chiral backbones or containing other catalytically competent groups (like e. g. the already mentioned selenoureas), should be accessible and we also think that the future will show that these catalysts also hold potential to facilitate novel transformations that are not possible with existing systems so far. Accordingly, the design and development of novel bifuncational Quatsalts will remain a rewarding task, at least in our opinion.

Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.