University of Groningen
Dynamic transfer of chirality in photoresponsive systems
Pizzolato, Stefano Fabrizio
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date: 2017
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Pizzolato, S. F. (2017). Dynamic transfer of chirality in photoresponsive systems: Applications of molecular photoswitches in catalysis. University of Groningen.
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
Chapter 4
Chapter 4
Studies towards a Trifunctionalized Molecular
Switch for Light-assisted Tandem Catalytic
Processes
This chapter describes the study towards a trifunctionalized molecular photoswitch based on an overcrowded alkene for light–assisted tandem catalytic processes. We proposed a two-step sequence of Morita–Baylis–Hillman (MBH) reaction and enamine catalyzed aldol reaction by merging two pairs of orthogonal bifunctional catalytic groups. Alternative designs, compared to those described in chapter 3, aimed to improve the catalytic activity in the MBH reaction and related attempted syntheses, are presented. Finally, screening of other transformations that could be mediated by the initially proposed photoswitchable catalysts design is reported.
4.1 Introduction
Catalysis is unarguably the most powerful tool to efficiently and effectively transform readily available building blocks into highly complex molecules and materials. Research focused for decades on the development of highly selective catalysts to fulfill any possible synthetic task via careful optimization to achieve high conversions and selectivities. While new synthetic methodologies are still under investigation, the novel field of switchable catalysis has recently emerged as a promising platform to further extend our control via more complex catalytic systems.1–6 Inspired by Nature, chemists are now crafting dynamically responsive catalysts whose activity and selectivity can be tuned or reversed by an external stimulus. Potential applications include the ability to control ‗one-pot‘ multi-component and multi-step synthetic processes, thus providing access to a variety of valuable products from a pool of building blocks depending on the order and type of stimuli provided. A brief selection of the most relevant systems for photoswitchable catalysis has already been presented in Chapter 3. For a comprehensive view of the field, the reader may refer to the recent reviews.1–6
Many systems have been demonstrated to achieve reversible control of the catalytic activity by external input, either by altering the reaction rate or by effectively switching ON/OFF the substrate conversion.7–15 More limited in number and variety are the examples through which dynamic chemoselectivity11,15–17 or stereoselectivity18–22 were accomplished. Our group largely established the potential of molecular motors23–
26
as versatile light-responsive central units for dynamic stereoselective catalysts,19,22,27,28 harnessing their unique tunable stereochemistry. Indeed, stimuli-responsive control of the activity and enantioselectivity displayed by chiral catalysts was achieved via dynamic conformational changes of a first generation molecular motor core equipped with two functional groups able to cooperatively accelerate a reaction. Such responsive organocatalysts (ROC1 and ROC2) and coordination ligands (RCL) were successfully applied, respectively, in the 1,4-addition of thiols to enones (Scheme 4.1a),19 the Henry reaction20 (Scheme 4.1b), and palladium-catalyzed enantioselective allylic substitution (Scheme 4.1c).22
Leigh and co-workers reported a multi-tasking rotaxane catalyst RC1 featuring a concealable secondary amine unit with which catalytic activity in an organocatalyzed reaction can be controlled by changes in pH (Scheme 4.2a).29 Effective control of the rate of 1,4-addition of an aliphatic thiol to trans-cinnamaldehyde was achieved via switchable iminium organocatalysis (Scheme 4.2b). Notably, further application of the same catalyst was accomplished due to the versatility of secondary amines in organocatalysis through other activation pathways, of which the non-protonated rotaxane (‗ON-state‘) displayed higher catalytic activity as a general trend.15 Effective β-functionalization of carbonyl compounds with S-nucleophiles (Scheme 4.2b) or C-nucleophiles (Scheme 4.2c) were reported through iminium activation and nucleophilic addition while substitution reactions were achieved via enamine catalysis (Scheme 4.2d). The rotaxane catalyst is even able to promote tandem iminium-enamine reaction sequences (Scheme 4.2e) and the Diels–Alder reaction of a dienal through a trienamine activation pathway (Scheme 4.2f). This concept of switchable catalyst was further extended to an asymmetric organocatalytic rotaxane that features an acyclic chiral secondary amine housed within a rotaxane framework.30 This system was able to control the rate of catalyzed asymmetric Michael addition of 1,3-diphenylpropan-1,3-dione to aliphatic α,β-unsaturated aldehydes. Good enantioselectivities (up to 93:7 er) were reported, however no dual stereocontrol could be afforded due to the fixed stereoinduction provided by the catalytic center.
Scheme 4.1. Stimuli-responsive control of catalytic activity and enantioselectivity achieved by dynamic
conformational changes of a bifunctional first generation molecular motor derivatives in organo- and metal-catalyzed transformations: a) 1,4-addition of thiol,19 b) Henry reaction20, c) palladium-catalyzed enantioselective allylic substitution.22
Scheme 4.2. Activation mode and scope of switchable rotaxane organocatalyst RC1 developed by Leigh
Noteworthy, Leigh and co-workers also developed a switchable rotaxane system featuring two different organocatalytic sites: a squaramide moiety and a dibenzylamine group (Scheme 4.3a).11 When the rotaxane is protonated, the macrocycle preferentially interacts with the ammonium unit revealing the squaramide unit, which can promote the conjugated addition of 1,3-diphenylpropan-1,3-dione to trans-β-nitrostyrene through hydrogen bonding catalysis (75% conversion after 18 h). In basic media, the macrocycle preferentially resides over the squaramide, revealing the secondary amine which promotes the Michael addition of 1,3-diphenylpropan-1,3-dione to crotonaldehyde via iminium ion catalysis (40% conversion after 40 h). In this way, the catalyst state controls which building blocks react together and which product is formed from a mixture of reactants (Scheme 4.3b).
Scheme 4.3. Activation mode and substrate-selective switchable rotaxane organocatalyst RC2 developed
by Leigh and co-workers.11
Despite the numerous approaches reported which are extensively reviewed in the literature,1–6 an artificial system based on a multitasking switchable catalyst or combination of multiple switchable catalysts able to control sequences of transformations has not yet been reported. The future impact and possible application of such challenging quest are definitely not easy to predict, as it may lead to the development of highly complex synthetic methodologies in a biomimetic-like fashion. An atom-efficient scenario achieved by performing several consecutive reactions in a ‗one-pot‘ (single reactor) is an attractive target. Greater economy of time and energy maximize resources and overall process simplicity, as well as decreasing materials loss from multiple iterations of reaction, workup and purification.31–35 Moreover, the use of light as a non-invasive external input allows for precise frequency, spatial and temporal control over functional groups response and chemical transformations, providing an artificial alternative to the feedback loops and trigger-induced effects typical of enzyme activity modulation.36
In this context, we engaged in the challenge of developing a prototype for a dynamically responsive multitasking catalyst,37–41 based on a second generation molecular motor core (Scheme 4.4). An interesting aspect of such a motor scaffold is the possibility of functionalizing the otherwise symmetrical lower half with two different catalytically active moieties (depicted as A and B, respectively), which could dynamically cooperate with the single functionality (C) located on the upper half in a trifunctionalized responsive core. Through this design two distinct bifunctional catalytic pairs could be alternatively activated, thus selectively and orthogonally promoting a two-step transformation sequence upon external input. X Y C B A h1 (P)-[AC] THI THI h2 (M)-[BC] R R S aryl X Y C A B X Y C B A (M)-[AC] (P)-[BC] aryl X Y C A B aryl aryl R S R h1 h2
Scheme 4.4. Proposed design of a trifunctionalized light- and heat-responsive organocatalyst for diastereo-
and enantioselective ‗one-pot‘ assisted tandem catalysis. The catalyst is envisioned to be switchable between four different states, each displaying a different combination of active cooperative catalytic pair (AC or BC) and helicity (P or M). In the scheme are displayed only two of the four possible products accessible by combining three starting components (depicted as different shapes) in a chemo- and enantioselective fashion (suggested handedness of the newly generated stereogenic centers indicated on the connecting bond). By triggering the proper catalyst states, both enantiomers of each diastereoisomer could be accessed.
Notably, each of these two cooperative catalytic pairs could also be selectively addressed in two pseudo-enantiomeric conformations (i.e. (P)-(AC) and (M)-(AC), (P)-(BC) and (M)-(BC), respectively) upon triggering the correct light and thermally induced isomerization processes. Indeed, the 4-step isomerization cycle featured by molecular motors provides access to two distinct catalyst configurations (E or Z), each of them displaying opposite helical chirality (P or M). Hence, we envision such a design as a feasible future route for developing the first responsive multi-tasking catalytic system capable of mediating ‗one-pot‘ multi-component transformations in a diastereo- and enantio-selective fashion. The complexity of the ideal target system is clearly grasped by listing the various critical requirements for such a design (Scheme 4.4).
a) The two catalytic steps must be promoted by two orthogonal distinct bifunctional catalytic pairs (AC and BC).
b) The shared catalytic moiety of the rotor (C) must be active in both bifunctional pockets. c) The product of the first catalytic step must be an active substrate for the second step.
d) The rate of the intermolecular cooperative catalysis, via interaction of two molecules of catalyst, should be negligible compared with the intramolecular cooperative mechanism, to efficiently suppress the undesired catalytic event.
e) In order to avoid further complexity of the catalyst design and restrict the provided stereoinduction to the sole dynamic chirality of the switch core, the catalytic moieties should not comprise additional stereogenic centers. The presence of additional chiral elements would in fact increase the number of possible diastereoisomeric forms of the catalyst, otherwise limited to the four isomers depicted in Scheme 4.4. Moreover, matched-mismatched effects caused by the different interaction between the distinct chirality of the switch core and the chirality of the functional groups A-B-C could be expected, thus further complicating the tandem catalysis development. f) The two steps of the tandem process should be performed in the same reaction conditions (solvent,
temperature, concentration, catalyst loading, etc.) in order to truly satisfy the requirements for ‗one-pot‘ multi-component transformation; alternatively, the modification of the conditions should not interfere with the system‘s performance.
g) The catalytic switch should be switched efficiently, reversibly and robustly, i.e. featuring high PSS ratios towards either the stable and metastable isomer, while displaying limited switching fatigue or decomposition.
h) The photo-generated metastable state should be highly thermally stable, i.e. feature a long half-life, thus retaining the desired configuration for extended time intervals (up to days) and non-cryogenic temperatures (up to 50 °C and above). This feature would ensure a limited variation of the catalyst mixture composition (ratio of stable vs. metastable) throughout the entire catalytic reaction, open the application to catalyzed processes characterized by long reaction times and allow higher working temperatures.
i) Substrates and products of the catalytic cycle should not be affected by photo-induced decomposition, nor interfere with the switching process of the catalyst, for instance via their own distinct light absorption or via quenching of the switch‘s excited state by intermolecular energy transfer.
j) The catalyst inhibition caused by substrates or products should be negligible. k) The catalyst must be chemically stable at the working catalysis conditions.
l) The synthesis of the trifunctionalized catalyst must be practically feasible and viable on a reasonably large laboratory scale, to eventually allow for proper screening of the catalysis conditions and optimization of the ultimate light-triggered tandem process.
m) If the product of the tandem catalyzed synthesis features newly formed stereogenic center(s), the catalyzed tandem sequence could be performed in an enantioselective fashion by use of a non-racemic catalyst mixture. Thanks to the inherent and dynamic helicity of the molecular motor design, asymmetric induction may be achieved with both stable and metastable isomers of the catalyst. However, asymmetric synthesis or resolution of the photoresponsive catalyst would be required.
n) The assisted tandem transformation may be composed of two processes generating distinct stereogenic centers. Noteworthy, the configuration of the secondly generated stereogenic center may be subject to stereospecific substrate control exerted by the previously generated stereogenic center rather than by stereoselective induction from the chiral catalyst.
Since the final goal of the project was to develop a catalytic molecular switch able to orthogonally mediate two distinct catalyzed transformations, we undertook a careful investigation of precedent literature for recent developments in catalyzed multicomponent ‗one-pot‘ reactions. ‗One-pot‘ reaction types include, but are not limited to: cascade (domino) processes;42 tandem catalysis (catalysts performing sequential transformations);43 multifunctional catalysts having more than one active site;44 dual catalyst systems where one catalyst enhances or alters the properties of the other catalytic cycle;45 and ‗one-pot‘ reactions involving isolated catalytic cycles, for example where a second catalyst is added after the first has
completed its transformation, or where the first catalyst is selectively deactivated later by addition of a second reagent.46 Mechanistically distinct from cascade or domino catalysis, in which a single catalytic transformation occurs sequentially, orthogonal tandem catalysis47 is a ‗one-pot‘ reaction in which sequential catalytic processes occur through two or more functionally distinct, and preferably non-interfering, catalytic cycles. Tandem catalysis has also been subcategorized into auto- and assisted-tandem catalytic cycles.43 Auto-tandem catalysis uses a single precatalyst to effect two or more mechanistically distinct catalytic cycles, typically by cooperative interaction between the various species in the system. In contrast, assisted tandem catalysis requires deliberate intervention in the system to switch between one catalytic cycle and another. Several examples of multi-catalyst promoted tandem or cascade reactions have been already described in literature,35 showcasing systems based on multiple metal-catalyzed transformations,43 organocatalytic domino reactions,48 and the combination of metal catalysts and organocatalysts.49,50 Marks and Lohr recently reported a perspective on orthogonal tandem catalysis, with particular focus on recent strategies to address catalyst incompatibility.51 They also highlighted the concept of thermodynamic leveraging by coupling multiple catalyst cycles to effect challenging transformations not observed in single-step processes, encouraging application of this technique to energetically unfavorable or demanding reactions. Noteworthy, this perspective mainly describes systems based on metal-catalyzed reactions, either using homogeneous or heterogeneous catalysts. Reviewed systems include examples applied to hydrocarbon upgrading via metal catalyzed olefin isomerization/metathesis, metal-catalyzed tandem arylation/heterocoupling, and enzyme- and acid-catalyzed glucose conversion to hydroxymethylfurfural. However, no examples of orthogonal tandem organocatalytically promoted transformations were described.
4.2
Results and discussion
4.2.1 DesignThe design of a switchable trifunctionalized catalytic system proposed hereto entails the reversible formation and activation of two distinctive catalytic pockets. It is worth mentioning that the ability to retain the switching properties should not be affected by individual electronic properties nor the cooperative interactions between the catalytic moieties. Despite the wide variety and solid background of metal-catalyzed processes, the formation of a metal-complex by cooperation of two ligating moieties (e.g. phosphines, amines, heterocycles) might have dramatic influence on the switching properties of molecular motor or switch. A preformed bidentate complex might be too stable to accommodate the isomerization of the overcrowded alkene bond, especially if each of the two ligating moieties was located on one half of the motor scaffold. It should be noted that previously reported examples of switchable coordination ligands based on first generation molecular motor for palladium-catalyzed transformation were changed upon irradiation of the responsive ligand not in presence of the metal source.22 However, in the case of labile metal-complexes the reversible coordination of the bidentate ligand may still permit the isomerization of mono- or non-coordinated species upon irradiation even in presence of the metal source, as demonstrated subsequently for an analogous responsive core displaying dynamic control of chirality and self-assembly of double-stranded helicates.28 On the other hand, a different molecular switch design based on a photo-responsive core of which one half fully contains a flexible bidentate ligand unit may constitute a promising approach (see Chapters 6 and 8).
We decided to focus our attention on the fast-growing field of organocatalysis (i.e. the catalysis with small organic molecules in the absence of metals or metal ions) with the prospect of designing the first molecular motor-based multi-catalyst based purely on organocatalyzed processes. A plethora of organocatalyzed transformations have been developed in the last decades,52 which witnessed the rise of numerous catalytic systems based for instance on enamine53 or iminium54 activated intermediates, phosphoric acids, N-heterocyclic carbenes55 or H-bond donors in asymmetric catalysis.56 We envisioned that small molecule H-bond donors in combination with Lewis-acid/base mediated catalysis could be implemented in our target system, while preserving its switching properties. This is also supported by our precedent successful
applications of first generation molecular motor based smart systems to achieve reversible stimuli-controlled catalytic transformation and anion binding.19,20,27 The Aldol reactions,57 conjugated additions,58 cycloadditions,59 Strecker reactions,60 Morita–Baylis–Hillman reactions,61 Mannich reactions,49 Henry reactions62 and nucleophilic additions to nitroolefins63 are only a few of the organocatalyzed transformations successfully developed so far.64 Several privileged catalytically active functional groups have been harnessed in effective organocatalysts, ranging from amines to phosphines, alcohols, thioureas, guanidines, amides, thiols, carboxylic acids, carbenes, the well-established proline and its related derivatives. While all these topics support our expectations to find multiple reactions which might be catalyzed by our goal multi-catalytic switches, it also generated doubts about their plausible selectivity and specificity.62 It may go without mentioning that all the reported organocatalyzed transformations harness the reactivity of particularly susceptible functional groups (e.g. imine, aldehyde, indole, α,β-unsaturated carbonyl group, keto-esters, etc.) already at mild reaction conditions, as opposed for example to more robust functional groups (e.g. aromatic halides, non-activated alkenes, non-activated allylic groups, etc.) converted via harsher metal-catalyzed processes. Conversion of a substrate into a product inherently entails a decrease in functional group reactivity (e.g. secondary amine, secondary alcohol, substituted heterocycles, non-activated carbonyl group, etc.). Subsequent derivatization would often be required to increase the intermediate reactivity towards a consecutive organocatalyzed process, unless a proper tandem or domino sequence is applied. However, it should be noted that most of the reported ‗one-pot‘ tandem organocatalyzed transformations in fact rely on an auto-tandem catalysis mechanism.43 Hannedouche and co-workers recently reported the first use of a multitask chiral ligand in an asymmetric assisted tandem catalysis protocol that successively combines a metal-catalyzed alkyne hydroamination followed by an asymmetric organocatalyzed Friedel–Crafts alkylation.65 To the best of our knowledge, no examples of fully organocatalyzed assisted tandem transformations have been reported to date.
In this context, we selected the Morita–Baylis–Hillman reaction (or MBH reaction, Scheme 4.5) as a starting point for the development of our multi-catalytic system. The classical MBH reaction is a carbon-carbon bond forming reaction yielding α-methylene-β-hydroxycarbon-carbonyl compounds by addition of α,β-unsaturated carbonyl compounds to aldehydes.
Scheme 4.5. General scheme of Morita–Baylis–Hillman reaction.
Instead of aldehydes, imines can also participate in the reaction if they are appropriately activated: such process is commonly referred to as the aza-Morita−Baylis−Hillman (aza-MBH) reaction. In either case, this reaction provides a densely functionalized product, yielding a carbonyl-derived allylic alcohol or secondary allylic amine upon addition to an aldehyde or an imine, respectively. Such characteristics had a strong influence during the design of our proposed tandem process, as it retains a potentially reactive α,β-unsaturated carbonyl motif in the product. Therefore, further functionalization could potentially be conducted via a second organocatalyzed process in a ‗one-pot‘ fashion. Most effective catalysts for MBH reactions are nucleophilic unhindered tertiary amines, such as DABCO and quinuclidine, or tertiary phosphines like tributylphosphine.66–68 In particular the specificity of tertiary amines as catalysts in the MBH reaction held good promise for our studies, as such functions may provide the orthogonality required for an efficient assisted tandem catalysis process achieved by a switchable multiple organocatalyst. Reaction conditions are often mild (temperature ranging from -20 °C to 40 °C), however reaction rates are notoriously low (reaction times from hours to weeks). The catalytic cycle consists of three steps: 1)
conjugate addition of the nucleophilic trigger catalyst to the β-position of the activated alkene motif; 2) nucleophilic addition of the α-position of the resultant zwitterionic adduct to the carbonyl or imine functionality of the electrophilic partner, also referred to the electrophilic quench; 3) proton transfer and elimination of the catalyst, thus restoring the α,β-unsaturated carbonyl functional group.69 According to the commonly established mechanism, the electrophilic quench is the rate determining step (second order in aldehyde and first order in nucleophilic catalyst and enone).70,71 Rate enhancement can be achieved by stabilizing the zwitterionic intermediate or by activating the aldehyde, for instance by a H-donor co-catalyst.72 In more elaborate designs, the reported catalyst system comprises both catalytic partners linked by a rigid chiral scaffold. The bifunctional catalysts based on the popular Cinchona alkaloid scaffold are a clear example.69 In summary, the MBH reaction features critical advantages for the current study: (i) the MBH products are flexible and multi-functionalized, retaining an inherent reactivity potentially exploitable in a tandem sequence; (ii) the recently reported methodologies often involve the use of a bifunctional organocatalytic systems; (iii) the MBH reaction is usually conducted under mild reaction conditions, which well suit the very slow thermal relaxation process displayed at room temperature by our chosen switch scaffolds (six-membered ring thiopyranyl upper half, five-membered ring fluorenyl lower half, see Chapter 2 and 3).
Recently, urea/thiourea-based bifunctional catalysts have emerged as powerful catalysts in a wide range of asymmetric transformations. 39,73–78 Their high activities as well as their selectivities were attributed to their ability of activating both electrophilic and nucleophilic centers of the reacting partners.40 Wang and co-workers reported the use of a chiral binaphthyl-derived amine-thiourea catalyst 1 for asymmetric MBH reaction (Scheme 4.6), proposing a synergy between the nucleophilic activation of the 2-cyclohexen-1-one via reversible conjugated addition of the tertiary amine and the dual H-bond stabilization of the generated enolate anion by the thiourea moiety.38,79
Scheme 4.6. Binaphthyl amine-thiourea catalyst 1 for enantioselective MBH reactions by Wang and
co-workers.79
Previously introduced, photoresponsive stereoselective catalysts based on a first generation molecular motor ROC1 and ROC2 feature the combination of DMAP and thiourea functional moieties (Figure 4.1).7 These precedents indicate that a thiourea group and an aromatic amine conjugated to the motor core may be well tolerated by our target trifunctionalized switch. Similarly, a tertiary aromatic amine substituent was successfully implemented in a nitro-amine disubstituted chiroptical molecular switch 2, which displayed efficient reversible photoswitching by use of UV and visible light (Figure 4.1).26 Inspired by these preceding results, we opted for a thiourea substituent as a hydrogen-donor moiety in the upper half and a basic dimethyl amine group in the lower half of molecular switches 3 and 4 to design the first bifunctional cooperative catalytic pair (see Chapter 3).
Figure 4.1. DMAP-thiourea substituted molecular motor ROC1-2 and nitro-amine substituted chiroptical
switch 2 previously developed in our group. Proposed design of bifunctional molecular switches 3-4 described in Chapter 3.
The second active site of our photoswitchable trifunctionalized catalyst was proposed to harness the cooperation of the shared thiourea moiety, linked to the upper half, together with a primary or secondary aliphatic amine, linked to the lower half. In the general mechanism, amine catalysts activate carbonyls by the formation of either an enamine or an iminium ion intermediate. Enamine formation raises the HOMO energy, increases nucleophilicity, and facilitates nucleophilic addition and substitutions reactions (Scheme 4.7a).53,80 Conversely, iminium ion formation increases the electrophilicity of the carbonyl carbon and lowers the LUMO energy.54 This allows access to pericyclic reactions and electrophilic addition reactions, particularly conjugate additions (Scheme 4.7b).
Scheme 4.7. Enamine and iminium ion activation of saturated and α,β-unsaturated carbonyls.
The obtained bifunctional cooperative catalyst could then be employed, for instance, in an aldol-type reaction with electrophiles (Scheme 4.8a) by activation of the carbonyl partner via enamine catalysis and hydrogen-bond activation of the electrophile provided by the upper thiourea substituent.38,39,53,58,81 Alternatively, the α,β-unsaturated carbonyl motif could also be activated via iminium catalysis for cycloaddition or nucleophilic 1,4-addition reactions (Scheme 4.8b).54
A cyclic aliphatic amine substituent could be considered a promising option in terms of catalytic activity and versatility for the design of a catalyst for aldol-type reactions. Scheme 4.8c illustrates the proposed design for bifunctional catalyst 5 featuring an imidazoline or imidazolidinone substituent directly attached by one of the nitrogen atoms to the fluorenyl lower-half. MacMillan and co-workers developed an efficient imidazolidinone-based catalyst for enantioselective aldol methodology.40
Scheme 4.8. a) General scheme of an enamine-mediated aldol condensation. b) General scheme of an
enamine-mediated conjugated addition of ketones to activated olefins.
Inspired by their work, we considered such a structure as a solid starting point for our bifunctional catalyst. First, such imidazolidin(on)e groups could simplify the synthesis through a straightforward Buchwald-Hartwig coupling with a halogen-substituted fluorenyl lower half. Second, it would also avoid the complication of generating a stereocenter within the bridging unit (vide supra). An alternative cyclic pyrrolidine/pyrrolidinone substituent would in fact be connected to the switch lower half by a chiral tertiary carbon, thus potentially resulting in a diastereomeric mixture of the target catalyst.
The concept design of a trifunctionalized photo-responsive organocatalyst for ‗one-pot‘ multi-step synthesis via Morita–Baylis–Hillman and subsequent enamine mediated aldol condensation is presented in Scheme 4.9. Merged together, the two catalysts 3 or 4 and 5 would constitute the first multi-catalytic molecular switch, able to perform a MBH-aldol reactions sequence in a ‗one pot‘ orthogonal tandem process upon light-assisted isomerization of the catalyst. When the catalyst is switched to the E-isomer, the cooperative activation by thiourea and tertiary amine groups is envisioned to promote the organocatalytic MBH reaction of enones and aldehydes to generate the intermediate MBH adducts. Upon photoisomerization, the catalyst could be switched to the Z-isomer, allowing the conversion of the MBH adduct to more complex final products, e.g. via thiourea and primary/secondary amine cooperative catalyzed aldol-type transformation or Michael 1,4-addition via iminium catalysis (not shown in scheme) upon activation of the second catalytic pair. S N HN S N H N organocatalytic Baylis-Hillman reaction O R' OH organocatalytic enamine alkylation reaction O R' El OH h1 h2 El N H N H S CF3 F3C S N H CF3 CF3 O NR3 H O NR3 H R' O H H H R' H O N R' OH R R N R' OH R R E tertiary amine + thiourea catalysis secondary aliphatic amine + thiourea catalysis O Thiourea Thiourea H H Thiourea aryl aryl E-isomer Z-isomer R HN R
Scheme 4.9. Proposed design of a trifunctional photo-responsive organocatalyst for ‗one-pot‘ light-assisted
tandem catalysis via MBH reaction (E-isomer, thiourea + tertiary amine) and subsequent enamine mediated aldol condensation (Z-isomer, thiourea + primary/secondary aliphatic amine).
4.2.2 Preliminary testing of bifunctional overcrowded alkenes as switchable catalysts
Morita–Baylis–Hillman reaction
Before starting the challenging synthesis of a multicatalytic molecular switch, we engaged a more systematic approach by testing the basic concept of a photo-activated reversible ‗ON/OFF‘ catalyst on a simpler prototype. The trifunctionalized catalytic switch design was split into the two corresponding catalytic bifunctional components. Therefore, we suggested two initial designs 3 and 4 for a catalytic molecular switch for MBH reactions (Figure 4.1). They feature a thiourea substituent as hydrogen-donor moiety in the upper half and a basic aromatic dimethylamine group in the lower half. In our previous study we showed that the combination of a 5-membered ring in the lower half (fluorene) with a sulphur containing 6-membered ring in the upper half (5,8-dimethylthiochromene and benzo[f]thiochromene) resulted in distinctive high energy activation barriers for the thermal relaxation step in the rotary cycle of the second generation molecular motors and consequently long half-lives of the corresponding metastable species. Moreover, by comparison of two structurally different upper halves we envisioned to investigate the influence of the distance between the two cooperative catalytic functionalities on the catalytic performance. Despite the literature claims by Wang and co-workers,79 our attempts to catalyze the model reaction between 2-cyclohexen-1-one and 3-phenylpropionaldehyde by either (E)-3, (Z)-3, (E)-4, (Z)-4 or the original literature catalyst 1 did not lead to any conversion to the desired MBH adduct as determined by GC-MS and 1H NMR spectroscopy analysis (Scheme 4.10).
Scheme 4.10. Attempted catalysis of the MBH reaction between 2-cyclohexen-1-one and
3-phenylpropionaldehyde using 1, stable isomers (E)-3, (Z)-3, (E)-4 or (Z)-4 as catalyst; no conversion to the desired Michael adduct was observed by 1H NMR. Conditions according to literature.79. Conversion monitored by GC-MS and 1H NMR spectroscopy analysis of crude mixture.
We postulate that these disappointing results are due to the limited catalytic activity of the dimethylaniline moiety, both in terms of the low nucleophilicity of the aryl amine and the structurally constrained nature of the tertiary amine within the catalyst‘s structure. In order to sustain our hypothesis, the influence of both nucleophilic and H-bond donor partners as separate co-catalyst units was addressed (Table 4.1). Notably, no conversion to the MBH product was observed upon use of DABCO (1,4-diazabicyclo[2.2.2]octane) or 6 in neat conditions (entries 1-2). When the two components were combined, only the reaction conducted without addition of solvent provided a moderate conversion to the expected product (entries 3-4), as proven by comparison with precedent reported 1H NMR spectral data. To further support the hypothesis of lower reactivity of the aromatic dimethylamine substituent, N,N‘-dimethylaniline (DMA) was tested alone and in combination with thiourea 6 resulting in no conversion in either cases (entries 5-6).
Table 4.1. Influence of Lewis-base (co-catalyst A) and thiourea 6 (co-catalyst B) in the MBH reaction
between 2-cyclohexen-1-one and 3-phenylpropionaldehyde.
Conjugated addition of 2,4-pentadione to trans-β-nitrostyrene
Compound 1 was also reported to catalyze the conjugated addition of 2,4-pentadione to trans-β-nitrostyrene.82 Similar to the original study on catalyst for MBH reaction by Wang,79 the proposed mechanism entails the acid-base cooperative activation by the bifunctional chiral catalyst of both reaction partners. 2,4-pentadione is coordinated in its nucleophilc enolic form by the dimethylamine group. The Michael acceptor trans-β-nitrostyrene is coordinated and activated by the thiourea group via double H-bond donation. Wang‘s catalyst 1 and switches (E)-3, (Z)-3, (E)-4 and (Z)-4 were eventually tested in such transformation by following the reported procedure. Similarly to the test results for the MBH reaction, no conversion towards the Michael addition product was detected in any of the cases as determined by GC-MS and 1H NMR spectroscopy analysis (Scheme 4.11).
Scheme 4.11. Literature reported82 and tested results for conjugated addition of 2,4-pentadione to trans-β-nitrostyrene using Wang‘s catalyst and bifunctional switches (E)-3, (Z)-3, (E)-4 and (Z)-4. Conversion monitored by GC-MS and 1H NMR spectroscopy analysis of crude mixture.
4.2.3 Attempted synthesis of alternative bifunctionalized switches
The initially proposed and subsequent alternative designs of a bifunctional switchable catalyst for MBH reaction are presented in Figure 4.1. In accordance to more commonly encountered functional groups featured by previously reported catalysts for MBH reactions, a small selection of alternative Lewis base motifs was proposed (Figure 4.2). The stronger nucleophilic character of aliphatic tertiary amines featured by 7, 8 and 9 and tertiary phosphines featured in 10 should ensure the desired higher catalytic activity.
Entrya Co-catalyst A (mol%) Co-catalyst B (mol%) Neat - Solvent Conversion (%)b
1 DABCO (20) / neat No conversion
2 / 6 (20) neat No conversion
3 DABCO (20) 6 (20) neat 60
4 DABCO (10) 6 (10) acetonitrile No conversion
5 DMA (20) / neat No conversion
6 DMA (20) 6 (20) neat No conversion
[a] Conditions: 2-cyclohexen-1-one (0.75 mmol, 3 equiv), 3-phenylpropionaldehyde (0.25 mmol, 1 equiv), neat or in acetonitrile (1 mL) as specified, room temperature, 4 d. [b] Conversion monitored by GC-MS and 1H NMR spectroscopy analysis of crude mixture.
Figure 4.2. Initial and alternative proposed designs of bifunctional switchable catalyst.
Scheme 4.12 illustrates the proposed retrosynthetic analysis of catalysts C and D. Similarly to 3 and 4, the more sensitive thiourea motif in the upper half is to be installed after the construction of the tetrasubstituted alkene and amination of the upper half via a Buchwald-Hartwig coupling.
Scheme 4.12. Retrosynthetic analysis of 7 and 8 starting from bromo-substituted diazospecies 12 and
tertiary amine-substituted thioketone 13.
The lower half coupling partner was synthesized from commercially available 2-carboxaldehyde-fluorene
14 (Scheme 4.13). Reductive amination of the carboxaldehyde moiety was conducted with a commercial
solution of dimethylamine (2M in MeOH) in presence of titanium(IV) isopropoxide and sodium borohydride to afford 15 (87%). Oxidation of the fluorene core to ketone 16 (83%) was then achieved under an atmosphere of air using the trialkyl ammonium salt Triton B as a phase transfer catalyst in pyridine. The conversion of fluorenone 16 into the reactive thioketone 17 was subsequently unveiled to be the weak link of the proposed synthetic route. Reaction with either P4S10 (Scheme 4.13a) or Lawesson‘s
reagent (Scheme 4.13b) resulted in rapid conversion of substrate 16, as observed by TLC analysis of the reaction mixture with disappearance of any spot other than at the baseline already after few minutes. However, no fraction resembling neither the substrate nor the product was collected upon flash column chromatography of the residual crude. By comparison with the synthesis of 2-(dimethylamino)-9H-fluorene-9-thione, the thionation of 16 appears to be detrimentally affected by the stronger basicity/nucleophilicity of the aliphatic tertiary amine substituent. TLC analysis of the reaction mixture showed a byproduct‘s spot (Rf < 0.05) at a very low elution speed in various mixtures of dichloromethane,
ethyl acetate and methanol. This may suggest the formation of a stable nitrogen-phosphorus adduct or salt rapidly generated in the tested conditions, thus preventing the conversion towards the desired thioketone. In a study conducted by Bergman and co-workers, the alternative thionating reagent P4S10·Py4 complex was
studied and employed for thionations of carbonyl functional groups in polar solvents such as acetonitrile and dimethyl sulfone, displaying excellent selectivity and substrate versatility.83 Its properties have been compared with the established Lawesson‘s reagent. Particularly interesting are the results from thionations in dimethyl sulfone at high temperatures (∼165–175 °C), at which Lawesson‘s reagent is inefficient due to rapid decomposition. The reported methodologies were successfully applied to a large variety of aliphatic and aromatic ketones, amides and ketamines, containing scaffolds such as lactams, nicotinamides, quinolones, indoles and oxindoles, to afford the corresponding thioketones and thioamides. Notably, no
aliphatic amine was included in the procedure scope, with the exception of glycine which afforded 2,5-piperazinedithione upon dimerization. Despite the lack of precedence, the thionation of 2-((dimethylamino)methyl)-9H-fluoren-9-one 16 with P4S10·Py4 complex in acetonitrile (Scheme 4.13c) and
dimethyl sulfone (Scheme 4.13d) were both attempted by following reported procedures. In either case, no trace of product was detected upon NMR analysis of the reaction crude. While the reaction in acetonitrile gave no visible conversion, reacting in dimethyl sulfone caused a sudden change from yellow to vivid purple color, a common indication of the presence of successfully thionated fluorenones. However, the reported workup step involves the use of boiling water to hydrolyze the excess of P4S10·Py4 complex, which
in our case rapidly caused the mixture to turn brown and may have also decomposed the expected product. When the procedure was subsequently repeated without the addition of water but using direct flash column chromatography of the solidified melted, no purple fraction was isolated. Eventually the herein proposed synthetic route towards 7, 8 and 9 via thionation of amine-substituted fluorenones and subsequent Barton-Kellogg coupling with diazo compound 12 was abandoned in favor of an alternative approach.
Scheme 4.13. Attempted synthetic route to tertiary aliphatic amine-substituted thioketone 17.
An alternative retrosynthetic analysis was proposed, as illustrated in Scheme 4.14. Secondary amine-substituted switches 7, 8 and 9 could potentially be obtained from a common dihalogenated intermediate
21, synthesized from the corresponding coupling partner bromo-diazochromane 12 and iodofluorenthione 22. Due to the expected higher reactivity of the iodo-substituted fluorenyl lower half, a plausible
chemoselectivity was envisioned upon lithiation and subsequent quenching with DMF to afford carboxaldehyde intermediate 20, or upon phosphination to yield phosphine intermediate 19, respectively. Similarly, quinuclidine-derived switch 18 may be obtained from intermediate 20 via condensation with 3-quinuclidone and subsequent reduction of the carbonyl group (for instance via a Wolff–Kishner-type reduction). Derivatization via Buchwald-Hartwig amination and subsequent introduction of the thiourea substituent would afford the target compounds 7, 8, 9 and 10.
Scheme 4.14. Retrosynthetic analysis of switches 7-10 starting from bromo-substituted diazothiopyran 12
and iodo-substituted thioketone 22.
The attempted synthesis of bromo-iodo-substituted alkenes 25 and 27 started with the generation of 2-iodofluoren-9-thione 22 upon reaction with Lawesson‘s reagent (Scheme 4.15). The unstable thioketone 22 was rapidly reacted in a Barton–Kellogg coupling with hydrazones 10 and 22 upon generation of the corresponding highly reactive diazo compounds with PIFA at low temperature. However, after sulfur extrusion with PPh3 or HMPT only small amounts of alkenes 25 and 27 were obtained by flash column
chromatography as inseparable mixtures of E- and Z-isomers. Separation of the stereoisomers could be achieved in a later stage of the synthesis. However, due to the discouraging results from these Barton Kellogg couplings and high catalyst loading often required in common organocatalytic transformations for final application of the target compounds as catalysts, the synthesis of 7-10 along this path was discontinued.
Scheme 4.15. Attempted synthesis of bromo-iodo-substituted overcrowded alkene 25 and 27 as precursors
4.2.4 Investigation of catalytic performance in alternative organocatalyzed transformations
The project was eventually redirected to a different approach in order to optimize the time left. As compounds 3 and 4 were already fully characterized, alternative reactions were tested to investigate their catalyst performance in different types of transformations other than the MBH. The proposed applications would harness either the basic and nucleophilic character of the dimethylamine substituent or its steric hindrance in combination with the hydrogen-bond donor nature of the thiourea motif. It should be noted that due to the very small quantities of 3 and 4 obtained, the screening tests described herein were performed by using Wang‘s catalyst 1 as a model. Due to its inherent similarity with Z-isomers of 3 and 4, the latter would have been tested once the successful reaction conditions were unveiled.
Alkylation of α-carboxypiperidones ethyl esters
The first investigated reaction was the synthesis of benzomorphan analogues by intramolecular Buchwald– Hartwig cyclization developed by Khartulyari and co-workers.84 In their study, the key bond formation was based on an intramolecular Buchwald–Hartwig enolate arylation reaction, to provide tricyclic benzomorphan derivatives. Thus, alkylation of α-carboxypiperidones ethyl esters with ortho-bromobenzyl bromides provides the necessary substrates. N-benzyl substituted piperidones were alkylated directly with substituted benzyl bromides (Scheme 4.16a). N-methyl substituted piperidone required alkylation via benzyl transfer by a pre-formed ammonium intermediate due to higher nucleophilicity of the nitrogen atom in the piperidone ring (Scheme 4.16b). Despite the elegant synthetic approach to pharmaceutically valuable targets, the described methodology provides only racemic products.
Scheme 4.16. Synthesis of benzomorphan analogues by intramolecular Buchwald–Hartwig cyclization
developed by Khartulyari and co-workers: a) N-benzyl protected piperidones directly alkylated with substituted benzyl bromides; b) N-methyl protected piperidone required alkylation via benzyl transfer from a pre-formed ammonium intermediate.84
It should be noted that the stereoselective event is the formation of the quaternary carbon via benzyl transfer to the enolate intermediate of the starting piperidones. Upon intramolecular Buchwald–Hartwig enolate arylation reaction, the tricyclic benzomorphane derivatives are subsequently constructed in a stereospecific fashion. We envisioned that a chiral inductor able to coordinate via hydrogen-bonding the prochiral enolate intermediate would give access to the corresponding enantioselective benzylation. Scheme 4.17 illustrates the proposed mechanism mediated by chiral benzyl ammonium derivatives of 1 in stoichiometric quantities. Similarly, using 3 or 4 as alkylating agents could allow achieving activity control upon photoswitching. The E-isomers would in fact characterized by a less efficient intramolecular transfer
due to the larger distance between the functional group compared with Z-isomers, possibly resulting in a lower alkylation rate. In alternative, a combination of activity and stereoselectivity control could be expected in the case of enantioenriched 3 or 4. Similarly to the dynamic control of stereoselectivity reported for catalyst ROC1 by Wang and co-workers,19 different enantioselectivity and reaction rate could be expected between the distinct of E- and Z-isomers of 3 or 4 as benzyl ammonium salts.
Scheme 4.17. Proposed asymmetric synthesis of benzomorphan analogues via asymmetric alkylation of
N-protected piperidones with a pre-formed benzyl ammonium chiral intermediate (derivatives of 1 and bifunctional switches 3 and 4) in stoichiometric quantities and subsequent palladium-catalyzed cyclization. The initial investigation was approached by using 1 as model compound. The tertiary amine moiety of 1 was successfully converted upon reaction with benzyl bromide to the corresponding ammonium salt 28 (Scheme 4.18a). Surprisingly, 28 was found to be particularly unreactive in the benzyl transfer to piperidone 29, despite the screening in toluene of temperatures from 50 °C up to reflux (Scheme 4.18b). Upon careful monitoring of the reaction by 1H NMR spectroscopy analysis over time, compound 28 was found to be slowly degraded to 1 and benzyl alcohol, as confirmed upon subsequent spiking of the sample with commercial benzyl alcohol for reference. The reaction was repeated according to the literature using the ammonium salt of dimethylaniline and benzyl bromide 31, which successfully yielded the expected alkylated piperidone 30 already at 50°C (Scheme 4.18c). While the aromatic dimethylamine substituent of
1 displayed the required nucleophilic character to generate the alkyl ammonium salt, 28 was found too
unreactive towards the alkyl transfer. It could be hypothesized the nucleophilicity of the naphthalenyl amine is higher than aniline, causing a greater stability of the corresponding ammonium salt. Alternatively, the high steric hindrance caused by the thiourea group may prevent instead the efficient approach of the piperidone enolate, thus impeding the alkylation. The bifunctional catalyst 1 was found again inactive under the tested conditions and our proposed enantioselective approach of the reported synthesis of benzomorphans derivatives was r unsuccessful.
Scheme 4.18. Preliminary test with 1. a) Synthesis of chiral benzyl transfer intermediate 28. b) Attempted
asymmetric alkylation of piperidone 30 by stoichiometric benzyl transfer with 29. c) Repeating literature procedure with dimethylamine-derived benzyl ammonium bromide.
Decarboxylative protonation of α-aminomalonates
Due to the unsatisfactory nucleophilic properties of the amine substituent of 1 and switches 3 and 4, we decided to test their catalytic performance as a Brønsted base. Asymmetric decarboxylative protonation of substituted aminomalonates in the presence of a chiral base is a synthetically convenient and straightforward route to synthesize a variety of natural and unnatural optically pure α-aminoacids. This synthetic methodology is based on the more general malonic acid synthesis where the chirality of the product can be generated during the enolate protonation step (Scheme 4.19a). Thiourea derived cinchona alkaloids promote the asymmetric decarboxylative protonation of cyclic, acyclic, or bicyclic α-aminomalonate hemiesters under mild and metal-free conditions to afford enantioenriched aminoesters in high yields and enantioselectivities. In particular, Rouden and co-workers reported the synthesis of both enantiomers of the aminoesters starting from racemic substrates using cinchona alkaloid 32 and its pseudoenantiomer derived from quinine (Scheme 4.19b).85] However, it requires stoichiometric amount of base (1 equiv) and long reaction times (7 d). In the proposed mechanism, the amine function could act as a chiral proton shuttle whereas the urea/thiourea group, a strong hydrogen-bond donor, would anchor the substrate to bring the chiral protonating agent in a close proximity to the prochiral enolate. An alternative mechanism may involve the interaction of carboxylate anion with thiourea that facilitates the decarboxylation step whereby protonation preferably occurs in a stereocontrolled fashion with the ammonium moiety of promoter 32. Despite the difference in basicity between an aliphatic amine featured by the cinchona alkaloids and the aromatic amine of 1, 3 and 4, we did envision the deprotonation process of aminomalonates not to be largely affected. Indeed, the subsequent decarboxylation could be triggered by the basic aniline motif, providing the target chiral amino-esters. Photoisomerization of switches 3 and 4 could give access to external control of activity or enantioselectivity depending on the selected catalyst isomer.
Scheme 4.19. a) Proposed mechanism of enantioselective decarboxylative protonation of
α-aminomalonates. b) Enantioselective decarboxylative protonation of α-aminomalonates mediated by thiourea cinchona alkaloid 32 developed by Rouden and co-workers.85
Preliminary tests on the decarboxylative protonation of 1-acetyl-2-(ethoxycarbonyl)piperidine-2-carboxylic acid 33 to ethyl 1-acetylpiperidine-2-carboxylate 34 are presented in Table 4.2. The tested conditions are according to the literature precedent.85 Both in absence or presence of DMA, conversion of the substrate was observed only at temperature above room temperature (entries 1-4). Upon addition of a substoichiometric amount of DMA, full conversion was still obtained only at higher temperature (entries 5-6). The use of the more basic DABCO as base gave full conversion already at rt (entries 7-8).
Table 4.2. Influence of temperature, base and thiourea catalyst in the decarboxylative protonation of
1-acetyl-2-(ethoxycarbonyl)piperidine-2-carboxylic acid.
Entrya Catalyst (mol%) Temperature (°C) Time (d) Conversion (%)b
1 / rt 4 No conversion
2 / 40 2 60
3 DMA (110) rt 4 No conversion
4 DMA (110) 40 2 Full conversion
5 DMA (20) rt 4 No conversion
6 DMA (20) 40 2 Full conversion
7 DABCO (110) rt 1 Full conversion
8 DABCO (110) 40 1 Full conversion
9 DMA (20) + 6 (20) rt 1 No conversion
10 DMA (20) + 6 (20) 40 1 25
11 1 (50) rt 4 No conversion
12c 1 (50) 40 1 45(racemic)
[a] Conditions: 33 (0.09 mmol) in THF (0.5 mL), catalyst loading, temperature and time as reported. [b] Conversion monitored by GC-MS and 1H NMR spectroscopy analysis of crude mixture. [c] Product was obtained as a racemic mixture as determined by
Compound 6 was then used in combination with DMA in substoichiometric amounts, to investigate the possibility of synergistic effects. Contrary to our expectations, low conversion was observed also at higher temperature (entries 9-10). Eventually, 1 was tested affording similar results to the blank reactions (entries 11-12). No enantiomeric excess was observed for the isolated product. Compared with the higher activity of DABCO, the large difference in pKb between the aliphatic amine in the cinchona alkaloid 32 and the aromatic amine of 1 might be the cause of lack of catalytic activity of the latter. Alternatively the acidity of the thiourea protons86 and the close proximity to the amine group might be responsible for its low basicity and poor catalyst activity, as suggested by the control experiment using DMA in presence of 6.
Alcoholysis of styrene oxides
Finally, we tested the binding properties of our thiourea-substituted switches. Numerous studies have been conducted on the supramolecular recognition of anions with ureas and thioureas derivatives.87 The success of this class of systems lies in their hydrogen bonding ability to a variety of reactants and intermediates in close similarity to those found in the active sites of enzymes. After the seminal reports by Wilcox88 and Hamilton,89,90 appropriate incorporation of the (thio)urea motif in acyclic, cyclic or polycyclic frameworks has become one of the prevailing strategies in the design of synthetic anion receptors.91 Recent advances in supramolecular recognition allowed chemists to design a wide range of synthetic receptors matching the requirements for inorganic and organic anion binding, such as halide anions, oxyanion, cyanide, nitronate, enolate anions and nitro-groups, with remarkable application as powerful organocatalysts.92,93 In particular, thiourea derivatives have been shown to effectively bind Y-shaped oxoanions such as carboxylates in polar solvents,94 achieving selective recognition of amino acids95 and synergistic effect in catalysis.85,93,96 Schreiner and co-workers reported a method for mild and regioselective alcoholysis of styrene oxides via cooperative Brønsted acid-type organocatalytic system comprised of mandelic acid 37 and N,N′-bis-[3,5-bis-(trifluoromethyl)phenyl]-thiourea 38 (Scheme 4.20a).97
Scheme 4.20. a) Method for mild and regioselective alcoholysis of styrene oxides via cooperative Brønsted
acid-type organocatalytic system comprising mandelic acid 37 and
N,N′-bis-[3,5-bis-p(trifluoromethyl)phenyl]-thiourea 38 developed by Schreiner and co-workers.97 b) Reported Proposed mechanism.
Various styrene oxides 35 are readily transformed upon addition of alcohols 36 into their corresponding β-alkoxy alcohols 39 in good to excellent yields. Simple aliphatic and sterically demanding, as well as unsaturated and acid-sensitive alcohols can be employed. The experimental findings suggested an H-bonding-mediated cooperative Brønsted-acid catalysis mechanism (Scheme 4.20b). It is likely that co-catalyst 38 coordinates to the acid 37 through double H-bonding, stabilizes the latter in the chelate-like cis-hydroxy conformation, and acidifies the secondary alcoholic proton via an additional intramolecular H-bond. The epoxide then is activated by a single-point hydrogen bond that facilitates regioselective nucleophilic attack of the alcohol at the benzylic position. The incipient oxonium ion reprotonates the mandelate ion and affords the β-alkoxy alcohol product. In this context, we envisioned switches 3 and 4 to potentially exhibit selective reversible cooperative effects with mandelic acid 37 upon photoisomerization. While E-isomers would provide effective binding of the acid and subsequent efficient catalytic activity, Z-isomers were expected to display either lower activity or strong asymmetric preference due to the steric hindrance generated by the amine substituent. Thioureas 1 and 6 were then tested beforehand to establish such potential. More precisely, 1 was used as always to mimic the Z-isomers of 3 and 4, while 6 might have provided insights on the use of E-isomers of 3 and 4, due to a reduce steric hindrance from the aromatic tertiary amine envisioned for the latter. The influence of thiourea and Lewis-base catalyst in the alcoholysis of styrene oxide 40 in presence of ethanol was investigated under conditions similar to those given for the work of Schreiner as presented in Table 4.3.
Table 4.3. Influence of thiourea and base in the catalyzed alcoholysis of styrene oxide.
Entrya Catalyst (mol%) Conversion (%)b Comment
1 / <5 Traces of 41
2 (R)-37 (1) 20 42 obtained, no traces of 41
3 (R)-37 (1) + DMA (4) No conversion /
4 DMA (4) No conversion /
5 (R)-37 (1) + 6 (1) 26 42 main product, 41 in traces
6 (R)-37 (1) + DMA (4) + 6 (1) No conversion /
7 (R)-37 (1) + 1 (1) 50 42 obtained, no traces of 41
8 (S)-37 (1) + 1 (1)) 30 42 obtained, no traces of 41
[a] Conditions: 40 (0.30 mmol), ethanol (1.80 mL), catalyst as reported, room temperature, 3 d. [b] Conversion monitored by GC-MS and 1H NMR spectroscopy analysis of crude mixture. [c] Product was obtained as a racemic mixture as determined by chiral HPLC analysis (AD-H, hept:2-propanol = 95:5, flow 0.5 mL/min, 40 °C, Rt 14.5 min and 15.0 min).
Surprisingly, no conversion to the expected product 41 was observed in any case. The blank reaction gave conversions to hydrolysed product 42 in traces (entry 1). The use of mandelic acid 37 alone only provided a higher conversion to 42 (entry 2). Notably, the addition of co-catalyst 6 resulted in no significant variation (entry 5). In any test where DMA was added, no conversion of 40 was detected (entries 3,4,6). It appears that an additional base suppresses the hydrolysis to sideproduct 42. The use of 1 in combination with either enantiomer of 37 resulted in conversion towards 42 to different extents. Noteworthy, Schreiner and co-workers reported that parallel reference experiments without 38, as well as experiments using 38 without acid co-catalyst 37 under identical reaction conditions, which showed no conversion to products 39.97 As opposed to 38, thioureas 1 and 6 feature a single 1,3-bis(trifluoromethyl)benzyl substituent. From the experimental evidence it appears that such structural difference causes a detrimental decrease in catalytic activity, possibly due to the large difference in acidity of and non-covalent interaction provided by the thiourea derivatives (compound, pKa in DMSO: 38, 8.5±0.1; 6, 12.1±0.1; 1, 10.72±0.02). Indeed, Schreiner
and co-workers reported a study on the acidities of popular (thio)urea organocatalysts in DMSO, which showed the incremental effect of trifluoromethyl groups (-CF3) on acidic strength as associated to
established catalytic activity in noncovalent organocatalysis.98 Due to complete lack of effective catalytic activity of 1 and 6 in such transformation, the investigation of the current project was eventually interrupted without testing 3 and 4. Eventually compounds (E)-3 and (Z)-3 were found to provide successful control of catalytic activity in the Michael addition reaction between (E)-3-bromo-β-nitrostyrene and 2,4-pentanedione upon irradiation of the stable isomers towards the corresponding metastable forms (see Chapter 3). Due to the insurmountable complications encountered during the development of an effective switchable organocatalyst based on a reversibly photo-responsive bifunctional overcrowded alkene, the venture of designing a switchable trifunctional catalyst for dynamic control of light-assisted tandem synthetic transformations was interrupted in favor of more practicable research proposals.
4.3
Conclusions
This chapter describes the study towards a trifunctionalized molecular photoswitch based on an overcrowded alkene for ‗one-pot‘ multi-catalytic systems. A detailed analysis of the requirements implied by such complex design is given. We proposed a two-step sequence of Morita-Baylis Hillman reaction and enamine catalyzed aldol or conjugated addition reaction catalyzed by merging two orthogonal bifunctional catalytic group pairs. As a preliminary investigation, we have presented the design and attempted synthesis of various photoresponsive bifunctionalized catalysts 7-10 for MBH reaction featuring combinations of thiourea and tertiary amine or phosphines groups. As opposed to compounds 3 and 4, the synthesis of compounds 7 failed due to incompatibility of the aliphatic tertiary amine with the thionation step required to obtain one of the coupling partners for the Barton-Kellogg reaction. An alternative route towards 7-10 from a common dihalogenated intermediate 27 was proposed. However, it was affected by particularly low yielding Barton–Kellogg couplings, after which the availability of small quantities of product would have complicated the subsequent stages of the study. Eventually, alternative reactions were tested to investigate the performance of 3 and 4 in mediating different types of transformations other than the MBH. More precisely, we explored their application in: conjugated addition of 2,4-pentadione to trans-β-nitrostyrene; alkylation of α-carboxypiperidones ethyl esters via benzyl transfer by a pre-formed ammonium intermediate; decarboxylative protonation of α-aminomalonates; and alcoholysis of styrene oxides. In all instances, the screening tests described herein were performed by using compound 1 as a model catalyst, due to its inherent similarity with the Z-isomers of 3 and 4. As opposed to our initial assumption, no activity was observed in any case. As demonstrated in this work, an aromatic amine substituent was shown to be a poorly active catalytic moiety. These studies provide valuable insight into the requirements for the design of more effective and complex trifunctionalized molecular switches, which may allow the photocontrol of catalyst activity and selectivity in multicomponent reactions. Key to the successful development of these future catalysts will be a deeper understanding of the compatibility of ancillary functional groups with the overcrowded alkene syntheses and the introduction of more active catalytic groups to ensure higher catalyst performance.
4.4 Acknowledgements
The author would like to thank Dr. B. S. L. Collins for her fundamental contribution to this work.
4.5 Experimental section
4.5.1 General methodsGeneral experimental details can be found in Chapter 3. Wang‘s catalyst 1 was synthesized starting from (R)-BINAM according to the reported procedure.79 P4S10·Py4
83
, N-benzyl-N,N-dimethylbenzenaminium bromide84, piperidones 2984 and 1-acetyl-2-(ethoxycarbonyl)piperidine-2-carboxylic acid 3399 were synthesized according the reported procedures. 2-iodo-9H-fluoren-9-one 23 was synthesized by Jort Robertus.100
4.5.2 Synthetic procedures
1-(9H-fluoren-2-yl)-N,N-dimethylmethanamine (15)
Compound 15 was prepared from commercially available fluorene-2-carboxaldehyde by a modified procedure previously reported.101
Titanium(IV) isopropoxide (2.65 mL, 10.30 mmol) was added dropwise to a commercially available 2M solution of dimethylamine in methanol (8.3 mL, 16.5 mmol, 3.2 equiv) followed by the addition of fluorene-2-carboxaldehyde 14 (1.0 g, 5.15 mmol). The reaction mixture was stirred at ambient temperature for 4 h, after which sodium borohydride (200 mg, 5.15 mmol, 1.0 equiv) was added and the resulting mixture was further stirred for another period of 1.5 h. The reaction was then quenched by the addition of water (3 mL), the resulting inorganic precipitate was filtered, washed with diethyl ether (20 mL) and the aqueous filtrate was extracted with diethyl ether (20 mL x 2). The combined organic extracts were dried on K2CO3, filtered and concentrated under reduced pressure. The
solid residue was recrystallized form EtOH : toluene (~10 mL) to yield 15 (1.10 g, 4.92 mmol, 95%) as light brown solid. m.p. 182-184 °C. 1H NMR (200 MHz, CDCl3) δ 7.76 (t, J = 7.9 Hz, 1H), 7.62–7.47 (m,
1H), 7.43–7.21 (m, 2H), 3.90 (s, 1H), 3.51 (s, 1H), 2.29 (s, 3H). 13C NMR (50 MHz, CDCl3) δ 143.4, 143.3, 141.6, 140.8, 137.5, 127.8, 126.7, 126.5, 125.8, 125.0, 119.8, 119.5, 64.6, 45.4, 36.8. HRMS (ESI, m/z): calcd for C16H18N [M+H] + : 224.1434, found: 224.1434. 2-((dimethylamino)methyl)-9H-fluoren-9-one (16)
A 100 mL two-necked round bottom flask fitted with a reflux condenser was charged successively with 1-(9H-fluoren-2-yl)-N,N-dimethylmethanamine 15 (970 mg, 4.34 mmol), pyridine (50 mL) and benzyltrimethylammonium hydroxide (40 wt% solution in EtOH, 0.20 mL, 0.1 equiv). An air inlet was then introduced through the septum and a stream of air was allowed to pass through the reaction mixture. The reaction mixture was then allowed to stir at rt for 3 h under this set-up. After this time the pyridine was removed under reduced pressure. The residue was then dissolved in CH2Cl2
(30 mL) and washed with water (3 x 30 mL), brine (30 mL), dried over MgSO4, filtered and concentrated
under reduced pressure. The crude reaction mixture was then purified by flash column chromatography (SiO2, NEt3 in pentane, gradient 10–25%) to provided the title compound
2-((dimethylamino)methyl)-9H-fluoren-9-one 16 (860 mg, 3.62 mmol, 83%) as yellow oil. 1H NMR (400 MHz, CDCl3) δ 7.66 (dt, J = 7.3,
0.9 Hz, 1H), 7.61 (t, J = 1.1 Hz, 1H), 7.54–7.45 (m, 4H), 7.29 (td, J = 7.0, 1.6 Hz, 1H), 3.45 (s, 2H), 2.28 (s, 7H). 13C NMR (100 MHz, CDCl3) δ 193.9, 144.4, 143.4, 140.4, 135.3, 134.6, 134.4, 134.3, 128.9, 125.1,
124.3, 120.2, 120.2, 63.8, 45.3. 13C NMR (100 MHz, CDCl3) δ 195.0, 151.2, 145.9, 135.7, 134.8, 134.2,
126.8, 124.1, 121.2, 118.9, 116.6, 108.4, 40.7. HRMS (ESI, m/z): calcd for C16H16NO [M+H] +
: 238.1226, found: 238.1228.
