MSc Chemistry
Master Track: Molecular science
Master Thesis
Covalent synthesis towards azobenzene functionalized
[2]rotaxanes
by
Nicol Heijtbrink
10580611
August 2019
48 EC
December 2018 – August 2019
First Examiner:
Second Examiner:
Prof. Jan van Maarseveen
Prof. Fred Brouwer
Daily Supervisor:
Martin Wanner
Van ‘t Hoff Institute for Molecular
Science, Synthetic Organic Chemistry
Abstract
0.1
Abstract
In this thesis work was done towards molecular machines. Rotaxanes are multi com-ponent mechanically interlocked structures which can prove useful motives for switches and molecular machinery. In a [2]rotaxane 2 components build up the molecule, one ring around one ’dumbbell’ shaped axle. These two components share a mechanical bond rather than a covalent one, meaning they are not chemically connected but the can not be separated without breaking at least one chemical bond. This means the components can have a motion relative to one another and this feature can be employed in molecular machinery. The objective was to functionalize a pre-existing [2]rotaxane system in such a way that switchable functionalities could be installed. The synthesis of this type of rotaxane is a template assisted covalent approach, which has proven successful for an ’unfunctionalized’ [2]rotaxane. To transform this system into a switchable molecular machine, a handle on the ring of the [2]rotaxane had to be installed which is prone to further functionalization. The first functional group to be installed as such a handle was chosen for an amine (NHBoc) group on the ring. However this proved to decompose under the hydrolysis conditions needed in the synthesis to release the ring from the template. Therefore another approach was taken and a methoxy functional group was installed on the ring. This research proves a successful template assisted synthesis via a capping mechanism to from rotaxanes with two methoxy groups on the ring fragment in 13% overall yield (see figure1). In this figure the blue part represents the ring of the rotaxane, the red part represents the template that is now part of the thread and the green parts represent the stoppers that prevent dethreading.
ii
0.2
Populair wetenschappelijke samenvatting
In dit onderzoek is er gewerkt om moleculaire machines te maken. De basis hiervan is een rotaxane aangezien dit uit meerdere mechanisch verbonden componenten bestaat. In een [2]rotaxane zijn 2 componenten aanwezig in het molecuul, een ring om een haltervormige as. De twee componenten zijn niet chemisch gebonden maar delen een mechanische binding. Dit betekent dat de ring en de as niet van elkaar gescheiden kunnen worden zonder minstens 1 chemsiche binding te breken ondanks dat ze geen chemische binding aan elkaar hebben. Hierdoor kunnen de componenten los van elkaar bewegen en deze eigenschap kan gebruikt worden in een moleculaire machine. Het doel van deze these was om een bekende [2]rotaxane zo te functionaliseren dat het een schakelbaar systeem wordt. De rotaxane wordt gesynthetiseerd via een covalente methode om een sjabloon heen. Deze methode is een bewezen techniek voor het maken van deze soort rotaxane. Echter was deze rotaxane niet gefunctionaliseerd, dus in dit onderzoek zal eerst de ring van de rotaxane gefunctionaliseerd worden met een hendel die achteraf verder omgezet kan worden naar een gewenste functionele groep. Eerst is er gekozen voor een amine (NHBoc) groep op de ring, echter bleek deze niet stabiel te zijn onder de hydrolyse condities die nodig zijn in de synthese om de ring van het sabloon los te maken. Daarom is er verder gekeken naar de methoxy groep op de ring, waaruit succesvol een [2]rotaxane is gesynthetiseerd, zie figuur 2. Dit is gelukt via een ’capping’-mechanisme om een sjabloon heen in 13% totale opbrengst. In het figuur is het rode gedeelte afkomstig van het sjabloon wat nu verwerkt zit in de draad, het blauwe gedeelte geeft de ring weer en de groene gedeeltes zijn de stoppers die er voor zorgen dat de ring niet van de draad glijdt.
Contents
Abstract i
0.1 Abstract . . . i
0.2 Populair wetenschappelijke samenvatting. . . ii
1 Introduction 1 1.1 Mechanically interlocked molecules . . . 1
1.2 Mechanisms for synthesis of mechanically interlocked molecules . . . 5
1.2.1 Supramolecular approaches . . . 5
1.2.2 Covalent mechanisms . . . 5
1.3 Aim of the thesis . . . 6
2 Results and Discussion 8 2.1 General approach . . . 8
2.2 Synthesis of the starting materials . . . 8
2.2.1 Terephtalic acid templates . . . 8
2.2.2 Stopper . . . 10
2.2.3 Functional groups on ring precursors . . . 11
2.2.3.1 NHBoc-ring . . . 11
2.2.3.2 Other functional groups via the Du↵-reaction. . . 12
2.2.3.3 Di↵erent substrates in the Du↵-reaction . . . 13
2.2.3.4 Methoxy-ring . . . 16
2.3 Assembling the rotaxanes . . . 17
2.3.1 Amine functionalized rotaxane . . . 18
2.3.2 Methoxy functionalized rotaxane . . . 20
2.4 Switches on the thread . . . 24
3 Conclusion 25 3.1 Conclusions on the synthesis of the starting materials . . . 25
3.2 Amine functionalized [2]rotaxane . . . 26
3.3 Methoxy functionalized [2]rotaxane . . . 27
4 Outlook 28
Contents iv
A Supplementary Information 32
A.1 General methods and materials . . . 32
A.2 Spectra . . . 47
A.2.1 Amine fundtionalized ”rotaxane” . . . 47
A.2.2 Methoxy fundtionalized [2]rotaxane . . . 48
Chapter 1
Introduction
1.1
Mechanically interlocked molecules
The subject of this thesis is a covalent synthesis towards functionalized rotaxanes. Ro-taxanes are special types of molecules because they consist of multiple separate compo-nents which do no not share a covalent chemical bond. Instead they are interlocked via a mechanical bond resembling two rings of a chain. These type of interlocked structures belong to a class of molecules called mechanically interlocked molecules (MIMs). This mechanical bond in MIMs di↵ers in properties from supramolecular assemblies but also from covalently bound molecules and therefore the term mechanical bond has been in-troduced to describe the relationship between the components of a MIM [1]. Figure 1.1
shows some of the most widely explored MIMs including rotaxanes and a catenane.
Figure 1.1: Di↵erent types of mechanically interlocked molecules
The most scientifically interesting feature of mechanically interlocked molecules such as rotaxanes and catenanes is that they are comprised of two or more components whose motion relative to each other can be controlled. A molecular shuttle based on a [2]ro-taxane for example, consists of one microcyclic wheel over an axle bearing two distinct
recognition sites. The ring can shuttle between the recognition sites and this motion is the principal mechanism behind many interlocked molecular switches or machines [2]. Synthesis of molecular machinery is aided by an understanding of the movement of pro-teins in biological systems because enzymes form their own molecular machines. For example directed sliding motion such a ring over an axle in a rotaxane is a mimic of myosin and kinesin mechanisms [3].
A catenane is a MIM that comprises two or more interlocked macrocycles. These intertwined rings can not be separated without breaking at least one covalent bond in one of the macrocycles, in resemblance to two rings in a chain. Catenation is not only a synthetic concept, it is also observed in DNA wherein catenanes are natural intermediates in the process of DNA replication of circular DNA molecules [4]. This naturally occurring phenomenon has lead scientists to synthesize similar structures in the form of ’simple’ catenanes, see figure 1.2. The first one on display has been synthesized by Stoddart et. al. via a template directed synthesis, based on noncovalent interactions [5]. The second example provided is a catanane by Sauvage et. al. which they synthesize via metal coordination. The macrocycles coordinate around a Copper(I) center to form the interlocked species [6]. The third example by Sanders et. al. is achieved via a di↵erent synthetic method called dynamic combinatorial chemistry [7]. This method makes use of a reversible assembly via hydrazone linkages of the building blocks around a receptor target molecule, see figure1.2(c).
(a) Catenane by Stoddart et. al. [5] (b) Catenane by Sauvage et. al. [6] (c) Catenane by Sanders et. al. [7]
Figure 1.2: Examples of catenanes, synthesized in various manners.
A [1]rotaxane, is also a synthetic analogue of a naturally occurring mechanically inter-locked molecule. A [1]rotaxane is a thread attached to a macrocyclic structure, which threads through this ring, much a like a lasso. No wonder that the natural peptides that have this shape are called lasso peptides, an example is provided in figure1.3. The shape of a lasso peptide is generally stabilized by steric interactions, preventing dethreading. However they can be further stabilized by the presence of disulfide bonds formed be-tween cysteine residues in the amino acid chain that builds up the peptide [8]. The lasso peptide family can be divided into three classes, depending on the number of disulfide
bonds present: class I has 4 cysteines and 2 disulfide bonds, class II has 0 cysteines and 0 disulfide bonds and class III has 2 cysteines forming 1 disulfide bond [9]. The lasso peptide shown in figure 1.3belongs to class II as there are no disulfide bonds present.
Figure 1.3: Structure of lasso peptide caulosegnin I as reported by Hegemann et. al. [8]
Comparing this lasso peptide to the general structure of a [1]rotaxane in figure 1.1, one can clearly see the structural similarities between the two. Though [1]rotaxanes are not specifically stabilized via disulfide bonds, they do make use of steric interaction to maintain their interlocked form. An example of a [1]rotaxane is shown in figure 1.4, which has been directly synthesized in a relatively high yield of 73% [10].
Figure 1.4: [1]Rotaxane synthesized by Li et. al. [10]
A useful motive for the macrocyclic part of a rotaxane can be -cyclodextrine (CD); a macrocycle consisting of 7 glucose units [11]. CD-based [1]rotaxanes have been synthe-sized, for example to stabilize radicals inside the CD cavity [12]. This can be ascribed to steric hindrance caused by the mechanical bond, which can reduce reactivity and so trapping otherwise highly reactive species such as radicals. Moreover, the switchable be-haviour of MIMs has also been investigated in CD-based [1]rotaxanes, see figure 1.5(a). Switchable behaviour of photoisomerizable azobenzene threaded, CD based [1]rotaxane
[13] and pseudorotaxane [14] has been demonstrated. In these systems light of a cer-tain wavelength switches the azobenzene moiety on the thread from the E isomer to the Z isomer and vice versa, so controlling the position of the ring on the thread. This concept will also be a base for the aim of this research. However it should be noted that azobenzene based switches are not the only possibility in terms of creating ring movement in a rotaxane. The [1]rotaxane shown in figure 1.5(b) can chemically switch between inclusion and exclusion of an ammonium/amine group by varying the pH of the solution [15]. Herein a di↵erent type of macrocycle was used as the ring; crown-ether instead of -cyclodextrine. In an acidic environment such as trifluoroacetic acid the amine is protonated to an ammonium what has a preference to be in the crown ether ring for stabilization via the electron donating oxygens in the macrocycle. Under basic conditions this amine is electron rich enough and therefore excluded by the ring.
(a) Light induced switchable azobenzene in a rotaxane
[13]
(b) Chemically switchable rotaxane [15]
Figure 1.5: Examples of switchable [1]rotaxanes
Besides the lasso peptide inspired [1]rotaxanes, higher order rotaxanes are also known. The increasing of the order of rotaxane has a linear relation to the increase of compo-nents. [2]Rotaxanes comprise of 2 components, usually one macrocyclic ring over an axle. Even higher order MIMs such as [3]rotaxanes are known. These consist of three com-ponents, usually two rings over an axle [16–18]. Also higher order catenanes are known, wherein the same concept applies, so a [3]catenane is build up of three components, all three being rings in this case [19–21].
Apart from the previously described MIMs there are also structures known as molecular knots [22–24] and molecular Borromean rings [25,26] which are outside the scope of this thesis.
1.2
Mechanisms for synthesis of mechanically interlocked
molecules
1.2.1 Supramolecular approaches
Most of the examples of synthetic MIMs in the previous section have been synthesized via supramolecular approaches. This means the assembly of the molecules makes use of reversible, non-covalent interactions between the components. Many supramolecular methods utilize self-assembly of the components around template, for example a tem-plate e↵ect around a metal ion [27]. The same principal can be applied in a metal free, anionic template [28]. The supramolecular nucleophile employed by Hubner et. al. showed affinity for lactam macrocycles and this characteristic was exploited in their syn-thesis of a [2]rotaxane. Another example of a template-directed molecular self-assembly process depends not only on ionic interactions but also on noncovalent bonding interac-tions such as ⇡ ⇡ stacking between an electron poor bipyridinium unit and an electron rich hydroquinol ring [29]. Finally hydrogen bonding is also included in the supramolec-ular approaches because of its usability in the synthesis of MIMs. For instance hydrogen bonds between crown ethers and ammonium ions can be applied to construct rotaxanes [30–32]. Yet hydrogen bonds between amides have also proven e↵ective as preorganiza-tional motives [33,34].
1.2.2 Covalent mechanisms
In contrast to all the already mentioned supramolecular mechanisms for the synthesis of rotaxanes and catenanes there are still some covalent synthetic pathways [35]. Co-valent synthesis has the advantage of not needing a preorganizational motive, which will be inevitably present in the final rotaxane (or catenane). The covalent strategies for the synthesis of rotaxanes can generally be divided in two categories; clipping and capping[36], see figure 1.6. Both will be applied in the ensuing research. The di↵erence between the two mechanisms is to the greatest extent the order of the reaction steps. Clipping is the term that describes the macrocyclization reaction that traps an axle of a rotaxane inside a macrocycle [37, 38]. This method is not exclusively used in covalent synthesis and can also be used in combination with supramolecular methods [39,40]. In each case the same foundation applies, first all the components are assembled, making sure the ring precursors, the thread and the stoppers are preorganized in the correct fashion. Thereafter the macrocycle forming the ring is closed leading to the mechanically interlocked rotaxane. Capping on the other hand relies on the fact that
Figure 1.6: Top: Schematic representation of a clipping type mechanism. Bottom: Schematic representation of a capping type mechanism.
the ring is already in place and closed prior to covalent capping with bulky stopper groups [41].
1.3
Aim of the thesis
The aim of this research is to synthesize a fluorescence switchable molecular machine. The foundation of the molecular machine will be a [2]rotaxane with a photoreceptive moiety on the macrocycle and fluorescent switches on the thread. This would create a stimuli-responsive rotaxane wherein light of a certain wavelength would activate the photoreceptive ring. The ring will then in turn flip the switch on the thread of the rotaxane to ”open the door” and let the ring slide to the other of the switch. The idea is that if there is a series of switches on the thread, the ring could walk stepwise over the thread. This allows tunable translational motion where relative positions of the interlocked components can be altered by the use of external stimuli. However in order to do this a synthetic route for functionalization of rotaxanes has to be developed. The chosen approach is a covalent template assisted synthesis. This has proven successful for the synthesis of the [2]rotaxane shown in figure1.7with R=tert butyl [35]. This has been achieved via a clipping type mechanism around a terephtalic acid based template. The research described in this thesis will primarily focus on the functionalization of the ring of the [2]rotaxane, transforming the R in figure 1.7 into an amine and a methoxy group. These will serve as handles for further functionalizing the rotaxane into a real molecular machine.
Previous work has been done in the group on a rotaxane with R = NHBoc, however research by a previous student (I. Bekaert) showed that preforming the synthesis via a capping type mechanism does not yield the desired rotaxane. Switching from a clipping to a capping mechanism hypothetically lead to wrong ring closing, leading to a non interlocked system. The first question to answer will be:
Figure 1.7: [2]Rotaxane synthesized by Steemers et. al. [35] with R = t-Bu
Can an amine-functionalized ring be incorporated the capping mechanism for the covalent synthesis of [2]rotaxanes?
Furthermore this research is focused on investigating other functional groups on the ring and so the next question that needs to be answered is
Can the ring in a [2]rotaxane, synthesized via covalent template assisted capping methods, be equipped with other functional groups on the R position in figure1.7?
Once this is answered the motion of the ring relative to the axle of the rotaxane will be investigated, however for this goal recognition sites have to be integrated in to the axle of the rotaxane. So the next question will be:
Can recognition sites, in the form of a photoisomerizable group such as azobenzene moi-eties be integrated in our [2]rotaxane system?
And finally the goal is to make molecular machines, so we need to answer the question: Can our current rotaxane system be functionalized in such a way that it can function as a fluorescence switchable molecular machine?
Chapter 2
Results and Discussion
2.1
General approach
In this research both the clipping and the capping mechanism have been used to build up ring-functionalized rotaxanes. Both mechanisms require certain building blocks; precur-sors for the ring, terephtalic acid core template and stoppers. For the two mechanisms these are then assembled in a di↵erent order. This research uses a template assisted approach for a covalent synthesis of [2]rotaxanes, so one of the key components in the synthetic route is the template. Because of earlier findings [35] a terephtalic acid based template has been used in either clipping and capping mechanisms. The other compo-nents needed to build up a rotaxane are a ring and an axle. The axle consists of a thread and stoppers on either side of the thread. However all of the starting materials are not commercially available and have to be synthesized, so the next section will describe the synthesis of all the starting materials.
2.2
Synthesis of the starting materials
2.2.1 Terephtalic acid templates
The template used for the synthesis of the [2]rotaxanes in this research is based on a terephtalic acid core. However, depending on the synthetic route one follows a di↵erent form of the template needs to be synthesized. However all templates make use of an activated ester in the form of a pentafluorophenol (PFP) ester, but they di↵er in the phenolic protecting groups. For a clipping mechanism the choice would be a core with thread precursor moieties attached to the template, as the framework of the rotaxane is build up before ring closing. The thread is build up via a copper catalyzed alkyne azide
cycloaddition reaction (CuAAC), therefore template A (figure 2.1) used in the clipping mechanism bears alkynes. However in the case of the capping method this is not suitable because the ring closing is one of the first steps and the alkyne groups in template A will interfere in the ring closing metathesis reaction. Therefore template B and C have been synthesized as precursors in a capping-type mechanism, which have an allyl and benzyl protecting groups respectively as easily removable groups on the phenolic OH’s (figures2.2and2.3). These protecting groups are necessary because in all cases the ring is linked to the template via a transesterification step wherein a phenolic OH of the ring precursor reacts with the activated ester on the template. If the phenolic OH’s on the template itself would not be protected unwanted side reactions could occur.
The synthesis of the thread attached template A is shown in figure2.1.
Figure 2.1: Synthesis of terephatic acid template A
The synthesis starts with a NCS-mediated oxidation to yield species 1. This intermediate is then alkylated using a mesylated 4-pentyl-1-ol, yielding the ester 2 with the alkyne thread precursors in place. 2 is sequentially converted to the activated PFP-ester via a saponification with potassium hydroxide, creating carboxylic acid groups which are then coupled to the pentafuorophenol with HBTU as a coupling reagent, yielding the final activated ester 4 which will from now on be referred to as template A.
Template B, which is now widely being used in the group as a template for the capping mechanism was synthesized according to the procedure displayed in figure 2.2. The synthesis is very similar to the synthesis of template A, however the alkylation step has been substituted for a allylation step using allylbromide.
Template C has been synthesized in our group by M. Wanner in the same manner as template B, however in stead of a protection of the phenol groups with allylbromide, protection was preformed using benzylbromide, see figure2.3. This template has inter-esting research value as it could optimize the synthesis. The advantage of this template is that the benzyl group does not interfere in the ring closing metathesis reaction, and is removed in the reduction step that comes after the Grubbs RCM to remove the internal alkenes in the ring.
Figure 2.3: Synthesis of template C
2.2.2 Stopper
In the design of the stopper it needs to be considered that the stopper is sufficiently large to prevent dethreading. The stopper should provide enough steric bulk so the ring can not slide o↵ the thread, so creating the mechanical bond between the components. Previous work has proven species 14 to be a suitable candidate [35]. This has been synthesized according to the procedure shown in figure 2.4.
Figure 2.4: Synthesis of the stopper
The first step in the stopper synthesis is lithiation of 4-tert-butylbromobenzene followed by quenching with dimethyl carbonate to give trityl alcohol 11 in 75% yield. Interme-diate 11 was reacted with malonic acid at 180°C to form the carboxylic acid 12 in 90% yield. This was reduced by BH3 dimethyl sulfide complex in 51% yield. The reason for this low yield is that during purification compound 11 was found in the reaction mixture, meaning that the malonic acid reaction was incomplete. Via the Mitsunobu reaction intermediate 13 was converted into the desired azide in 75% yield.
2.2.3 Functional groups on ring precursors
The original inspiration for this research was a [2]rotaxane with a tert-butyl group on the aromatic part of the ring, however this group is not the most suitable candidate for further functionalization, so a di↵erent functional group has to be installed in that position. This research focuses on the possibility of installing various functional groups on the ring which are prone to further conversion to for example fluorophores. As carbon to carbon or carbon to hydrogen bonds are stable ones, and not so easily functionalized the tert-butyl group could be best substituted for side groups based on nitrogen, oxygen or halogens.
2.2.3.1 NHBoc-ring
Previous research in the Synthetic Organic Chemistry group has shown a promising foundation to synthesize a ring with a tert-butyloxycarbonyl (Boc) protected amine group. The nitro group fucntionalized ring fragment 16 as shown in figure2.5had been previously synthesized according to figure 2.5. This was converted in two steps to the desired amine ring fragment 18. The first step was a reduction from the nitro group to a primary amine using SnCl2in quantitative yield. This aniline type compound is not a very stable intermediate so it had to be protected with a Boc-group as rapidly as possible. Removal of the tin salts formed in the reduction step proved to be a bit challenging and time consuming, possibly leading to decomposition of 17 prior to protection. This could be a explanation for the low yield of the Boc-protection step, which is usually a robust and high yielding reaction, but in this case only proceeded with 56% yield, even if a excess up to 5 equivalents of Boc anhydride was used.
2.2.3.2 Other functional groups via the Du↵-reaction
Not only the amine as functional group was analyzed in this thesis, also other functional groups have been tried to be installed on this ring fragment such as hydrogen, iodine, acetylamine and methoxy. The synthetic route for these precursors is di↵erent than for the NHBoc functionalized ring and starts of with a Du↵ reaction on a (functionalized) phenol. The mechanism of the Du↵ is shown in figure2.6[42]. It starts with a protona-tion of hexametylenetetramine (HMTA) by trifluoroacetic acid (TFA), forming a imine intermediate which then reacts with the phenol in a Mannich type fashion. This is the selectivity determining step, because as is visible the selectivity for this reaction is ortho to the phenolic OH group. It could also be para, however in most substrates the para position is blocked by another substituent leading to only ortho-substitution. A proton shift rearrangement, subsequent protonation, imine formation via the loss of ammonia and finally the last rearrangment leads to imine intermediate 25 which releases an alde-hyde product 26 after hydrolysis. For the purposes of this research the reaction has to proceed on both ortho positions in the molecule in order to form the final ring fragment, so the whole mechanism also proceeds once more to finally release a di-aldehyde product [43,44].
Figure 2.6: Mechanism of the Du↵ reaction
It is clear from this mechanism is that one HMTA molecule is sacrificed for each aldehyde that is formed, so two per di-aldehyde. Also visible is that the reaction relies on electron donation from the phenol group in order to proceed. So with one electron withdrawing group such as an aldehyde formed in the first step, the second reaction on the other side of the molecule will probably be harder to accomplish.
2.2.3.3 Di↵erent substrates in the Du↵-reaction
Various substrates were attempted in the du↵ reaction as it was desired to try and synthesize rotaxanes with rings with di↵erent functional groups. This kind of mono-and diformylations have been described in literature for a wide range of para substituted phenols [45, 46], so it was tried for a selection of desired starting materials for this synthesis: phenol, 4-iodophenol, paracetamol and 4-methoxyphenol. The first one to attempt was phenol, having a free hydrogen on the ring could be desirable for further functionalization via aromatic substitution or C-H activation, see figure 2.7. The idea was to add 2 equivalents of HMTA so no triple formylation would occur, however on this substrate the desired selectivity was not achieved under the tested conditions, see table 2.1. The position para to the phenolic OH is not blocked in this substrate so substitution on this position was also observed. In retrospect this was to be expected as this was described in literature found later [47]. However a Chinese patent claimed to have achieved selective formation of compound 28 [48], but having attempted the reaction many times under many reaction conditions it can only be concluded that they unknowingly were also working with the wrong product. In conclusion under various conditions mixtures of products were observed. The products were only separable by column chromatography and the desired product was not formed, so phenol is not a suitable substrate for further reaction.
Figure 2.7: Du↵ reaction on phenol
Reaction time Reaction temperature Eq. HMTA Conversion 27 28 29 30
1 day 130°C 2.2 28% 100% - -
-1 day 100°C 2.0 86 % 4% - 61% 35%
2 days 110°C 2.0 100 % 25% - 62% 13%
Table 2.1: Results from the Du↵ reaction on phenol
The next substrate was 4-iodophenol. By now it could be concluded that the para position has to be blocked to prevent unwanted substitution on that position. Halogens are useful groups for further functionalization, and because a coworker in the group was already working on 4-bromophenol it would be interesting to also try and synthesize the iodo-derivative and see the similarities and/or di↵erences between the two compounds. The 4-bromophenol has proven successful [46] and similar reactivity was expected for 4-iodophenol. However under the conditions required for a Du↵ reaction a variety of
products was observed, see figure 2.8. The first imminent observation was the loss of iodine, as this had sublimated on the reflux condenser after a night of reacting. Furthermore the mass of a product with 2 iodines was detected during analysis. 1H-NMR also showed a desymmetrized spectrum with two aldehydes, indicating formation of a side product like compound 33. In the bromide analogue this has not been described so this finding was surprising, but could also explain why this substrate is not described in literature. The better leaving group ability of iodine can be explained by the larger size of the atom and higher electronegativity of iodine compared to bromide causing this unexpected side reaction. Conversion in this reaction was low, and a mixture of undesired products was formed, see table2.2. So in conclusion this substrate is also not suitable for this reaction.
Figure 2.8: Du↵ reaction on 4-iodophenol
Reaction time Reaction temperature Eq. HMTA Coversion 31 32 or 33 34
1 day 110°C 8.0 28% - - 100%
1 day 110°C 2.0 37% - 40% 60%
Table 2.2: Results from the Du↵ reaction on 4-iodophenol
Paracetamol was also screened as a substrate for this reaction because amines could be interesting handles for further functionalization. The acetyl amine group in paracetamol could have either similar or di↵erent properties as the amine described in the previous section 2.2.3.1, either way interesting to research. Furthermore the starting material is readily available so it is worth investigating. The para position is now blocked, and the acetyl amine is not expected to show the same reactivity as the iodine derivative because it has worse leaving group ability. This substrate showed to be less reactive than substrates bearing alkyl groups on the 4 position [46]. Initially only mono formylated product 35 was observed. Increasing the temperature, equivalents of HMTA and reaction time pushed the substrate somewhat towards a di-aldehyde product 36, see table 2.3. However the yield was not high, the conditions were harsh and undesired product was still formed. The unwanted mono-aldehyde product could only be removed from the mixture by column chromatography, so taking in to consideration all the drawbacks it was concluded that this substrate was also not optimal.
The last substrate in this research to be considered is paramethoxyphenol, this is the only substrate of this research that has been reported, Lindoy et. al. reported a 42% yield
Figure 2.9: Du↵ reaction on paracetamol
Reaction time Reaction temperature Eq. HMTA Coversion 35 36
3 days 110°C 2.2 30% 100%
-1 day 110°C 3.0 64% 83% 17%
3 days 130°C 8.0 27% 50% 50%
Table 2.3: Results from the Du↵ reaction on paracetamol
for 2,6-Diformyl-4-methoxyphenol [46]. The methoxy group on this substrate could also prove to be a useful handle on the ring of a rotaxane for further functionalization. The para position is again blocked and it is not expected that the methoxy group will be lost from the substrate under the reaction conditions. However the remaining question is if it will be active enough in the double Du↵ reaction. Mono- and di-substituted products were observed (figure2.10) and increasing the harshness of the reaction conditions drove the product ratio towards the desired di-aldehyde product, see table2.4. However even after 3 days at 130 °C and 8 equivalents of HMTA, still 20% mono substituted product remained. Again the two products are only seperable via column chromatography. It is very well possible that increasing the temperature, reaction time and/or equivalents of HMTA could push the substrate on form only the double Du↵ product, but this was not investigated further. Optimization could be done to find the ideal conditions wherein the reaction proceeds in a clean and selective fashion.
Figure 2.10: Du↵ reaction on paramethoxyphenol
Reaction time Reaction temperature Eq. HMTA Coversion 37 38
3 days 110°C 2.0 43% 63% 37%
1 day 110°C 2.0 38% 44% 56%
3 days 130°C 8.0 76% 80% 20%
2.2.3.4 Methoxy-ring
From the intensive study conducted on the substrates and conditions of the Du↵ reac-tion, it could be concluded that the Du↵ reaction is not the ideal starting point for the synthesis of the ring precursor as this method is generally not clean and high-yielding. Therefore this was changed to a 2 step highly scalable literature procedure of hydrox-ylation using paraformaldehyde and oxidation facilitated by manganese dioxide [49] as shown in figure 2.11. MnO2 oxidized 39 to the same product as the Du↵ reaction 37 [50]. Even though the yield is low, no side products are formed and no column chro-matography is required for purification. The low yield of the oxidation reaction could be explained by the sticking of the product to the manganese. The MnO2 is removed from the reaction mixture via filtration, and filtrate contains the product as a yellow solution. Washing the residue with more solvent, warm ethyl acetate mixed with chloroform, re-leases more product from the residue, eventually reaching a yield of 37% for this step. This di-aldehyde product 37 can now be used as a substrate for a Grignard reaction [51] with 6-bromohexene to attach the alkane chains in 82% yield and finally a benzylic alcohol reduction with Et3SiH and BF3· Et2O yields the final methoxy functionalized ring fragment 41 in 82%.
2.3
Assembling the rotaxanes
Now having all the necessary building blocks at hand, the assembly of the [2]rotaxanes can start. In this research it was attempted to synthesize two di↵erent rotaxanes, one with a boc-protected amine group on the ring as a functional group and the second one with a methoxy group. Both have good prospects for further functionalization, but they were synthesized via di↵erent mechanisms. The NHBoc rotaxane was synthesized via a clipping type mechanism wherein all the components; so the ring precursors, the thread and the stoppers, were all assembled on the template prior to ring closing. In this mechanism the ring closing is the penultimate step before the release of the ring via hydrolysis. However the second rotaxane with methoxy groups on the ring was synthesized via a capping mechanism, meaning that the ring closing around the template is the first step and addition of the thread and stopper follow. This is considered a more versatile route because the thread and stoppers can be functionalized in a later stage in the synthesis.
2.3.1 Amine functionalized rotaxane
Figure 2.12: Total synthesis of the amine functionalized rotaxane
For the synthesis of the NHBoc functionalized rotaxane it was chosen to follow a clip-ping mechanism. Previous (unpublished) work in the research group has shown that
the capping approach was unsuccessful when this amine functionalized ring is used. De-threading was observed and it has been hypothesized that the ring closed on the same side of the thread, leading to no mechanical bond after hydrolysis of the template. The idea is that when using the clipping mechanism the steric bulk is increased before ring closing, forcing the ring around the thread in stead of outside of the thread. The whole synthesis is displayed in figure 2.12and will now be discussed step by step.
The first step in this mechanism is a transesterification of the pentaflourophenol (PFP) ester template A and NHBoc functionalized ring fragment 18. Cesium carbon-ate is used in this reaction as a base and promotes the transesterification. Wcarbon-ater can hydrolyse the PFP-esters of the starting material so 4 ˚A activated molecular sieves are added to prevent this even though the reaction is preformed in dry acetonitrle and under a flow of nitrogen. This yields the transesterification product 42 in 96% based on the terephtalic acid template. As the newly formed compound already carries the precursors to the thread, a copper catalysed azide-alkyne cycloaddition (CuAAC) ’click reaction’ follows directly. Also under dry conditions (to not poison the catalyst) this reaction yields 88% of compound 44. The CuAAC reaction has already previously been proven a useful reaction for installing functionality in MIMs [52]. Now all the groups have been installed all that remains is closing of the ring via ruthenium catalyzed metathesis. As briefly mentioned before this is done using a Grubbs II catalyst in an intramolecular metathesis reaction of the alkene ends of the ring precursor. This yields the ring with a mixture of E/Z isomers of the internal alkenes in 93%. This technique is also known to be used in the assembly of MIMs [21]. However, mostly for spectroscopic reasons, these double bonds are reduced using hydrogen and palladium on carbon yielding 46 with 78% yield. If this would not be done nuclear magnetic resonance (NMR) spectroscopy would show E and Z isomers of this double bond in ring making the spectra less clear and harder to interpret. In this stage the ring and axle are still covalently bound in stead of mechanically interlocked, making this a pre-rotaxane and not an actual rotaxane. In order to cleave the chemical bond holding the components together the ester bonds are hydrolyzed to theoretically yield the [2]rotaxane 47. In reality this did not occur, however not due to wrong ring closing as was previously hypothesized. It was discovered that the amine functionalized ring is not stable under the highly basic conditions needed for the hydrolysis, leading to decomposition of the ring fragment. In mass analysis the mass of the thread with the stoppers was observed but the ring was missing, see sup-plementary information. A test reaction using only the ring fragment 18 as starting material in the tesser’s base conditions that are used in prerotaxane hyrdolysis showed decomposition of the starting material in as little as two hours at room temperature on TLC. From this it can be concluded that 1 night under these basic conditions at ele-vated temperature, 50°C, the ring decomposed, and no actual dethreading occurs as was
previously hypothesized. Furthermore, the ring was not found in the reaction mixture analysis, supporting this finding. No further research has been done into the decomposi-tion product of this reacdecomposi-tion because purificadecomposi-tion was troublesome; baseline spots on the TLC could not be purified using column chromatography and 1H-NMR was inconclusive on the matter. However it is likely that the decomposition starts with deprotonation of the amine, and the ring is cleaved from there. In an attempt to increase the stability of the amine methylating the secondary amine to a tertiary amine using methyliodide was tried, see figure 2.13. 1H-NMR of the reaction product looked promising, however the prerotaxanes are quite large so in 1H-NMR they give many broad peaks and it was no definitive proof that the methylation worked. Therefore mass analysis was preformed but the correct mass was not observed.
Figure 2.13: Methylation of the ring in the amine functionized [2]rotaxane
The choice for the NHBoc group on the ring was made because it could provide a useful handle for further functionalization. However other functional groups could have the same purpose and could possibly be more stable under the applied conditions. That is why the rest of research is focused on the ring bearing a methyl protected phenol group (referred to as methoxy in the rest of this thesis).
2.3.2 Methoxy functionalized rotaxane
The next [2]rotaxane that was synthesized is the one with methoxy functional groups on the ring. The capping mechanism is considered a more versatile route for installing functional groups in the thread or on the stoppers, or even making higher order rotax-anes. There is no reason to believe the capping mechanism would not work for this type of rotaxane so it was chosen to make the rotaxane via this mechanism, see figure 2.14.
Figure 2.14: Total synthesis of the methoxy functionalized rotaxane
There are many similar steps to the clipping mechanism described in the previous section about the NHBoc-functionalized rotaxane, however there are a few di↵erences. The first one is the choice of template, for this mechanism template A is not suitable, but template
B and template C can be used, see figure2.2and2.3. The first step in the assembly is still a transesterification reaction on either template B or C. From this point the synthetic routes between the templates diverges. In our research group the template that is widely employed for a capping type mechanism for the synthesis of rotaxanes is template B. This needs to be deallylated at this stage because no ring closing metathesis reaction can be performed with the allyl protecting groups still in place as these will interfere, forming unwanted side products. So a palladium catalyzed deallylation reaction proceeded in 96% yield to yield intermediate 51, that is now ready for RCM. The isolated yield of the Grubbs II catalyzed RCM was only 48% because it is quite insoluble. For this reason it was challenging to purify it via column chromatography as the product does not dissolve well in eluent, and this is believed to be the main cause of the low yield. If another method of purification would be used it is likely that the yield can be improved. Further research can look into trituration or crystallization as purification methods to increase the yield of the RCM. Hydrogenation of compound 53 proceeded in a nice 95% yield, and this compound now converges paths with the product made from template C. However this product is synthesized in a slightly di↵erent manner from template C, see figure 2.15. In this case there are no allyl protecting groups, but benzylic ones. These do not interfere in RCM so there is no need for deprotection and one can continue immediately with the Grubbs metathesis after transesterification, which proceeded in 58% yield to form 57. The following hydrogenation step that reduces the olefin bonds in the ring also cleaves the benzyl groups, deprotecting the phenolic OH’s of the template, yielding the same intermediate compound 53. Likewise the yield is low, only 29%, probably due to solubility reasons and arduous purification.
Next in the synthetic route is the alkylation, which was preformed using pent-4-yn-1-yl methanesulfonate in the presence of base to yield compound 54 in 55%. This mesylate reagent is not commercially available due to limited stability so it has to be prepared fresh whenever this step is repeated, adding inherently one extra step to the (already long) total synthesis of a rotaxane. To overleap this, another alkylating agent that is commercially available was tested, the chloride analogue in figure2.16. However this did not seem to generate the desired product on NMR or TLC, so there was continued with the mesylate. Now having installed the alkynes and having azide stoppers in hand the copper catalysed azide-alkyne click reaction could be carried out. This yielded pre-rotaxane 55 in 89% yield. Finally hydrolysis of this intermediate released the [2]pre-rotaxane 56 in 95% yield.
Figure 2.15: Synthesis of methoxy rotaxane via benzylic template
2.4
Switches on the thread
The final goal of this research is to work towards a fluorescence switchable rotaxane to form a sequential molecular switch that can be operated via light of a certain wavelength. Unfortunately this came to fall outside of the scope of this project due to the limited time of the project. A switchable moiety needs to be installed in the thread of the rotaxane, for example in the form of an azobenzene. These groups are known to isomerize upon irradiation [53, 54] creating a gate that can open and close, trapping the ring on one side of the thread. The plan was to attach fluorophores to the now functionalized ring fragment, which could be absorb light, and emit it to irradiate the nearby photo-switchable group, so opening the gate close to the ring. However neither functionalities have been installed in the rotaxane yet. A good point in the synthesis to install the switches in the thread could be after the ring closing and hydrogenation, for example on 53 in figure 2.14. Likewise the fluorophores need to be installed on the ring. A plausible fluorophore is for example a BODIPY group (4,4-Difluoro-4-bora-3a,4a-diaza-s-indacene), which is known to strongly absorb UV radiation and emit relatively sharp fluorescence peaks with high quantum yields [55]. This chromophore could be installed in various stages of the synthesis. It might be wise to install these in one of the final stages of the rotaxane synthesis so they will not interfere in any of the earlier steps. Pre-rotaxane species 55 (in figure 2.14) could be a reasonable molecule for further functionalization so the BODIPY’s will be installed just prior to hydrolysis. It could of course also be investigated if the fluorophore can be installed in an earlier stage, for example in species 53, before capping with a thread and stoppers. However the methoxy group can not be functionalized directly, it should be deprotected from the methyl group to generate the free phenolic OH which is much more susceptible to various reactions.
Chapter 3
Conclusion
3.1
Conclusions on the synthesis of the starting materials
First of all a few conclusions can be drawn on the starting materials needed to assemble rotaxanes . Di↵erent templates can be synthezised following the same synthetic route, di↵erentiating in the side group on the phenolic OH’s to create a library of templates that can be employed for di↵erent purposes. Introduction of template C, and proof that it works in one simultaneous step to reduce the double bonds in the ring and remove the benzylic protecting group limits the steps in the total synthesis by one. The yields need to be optimized, which is most likely an improvement on the workup and purification method due to the low solubility of compound 53.
Furthermore a stopper was synthesized in a highly reproducible manner, the only point of attention is to make sure that step involving malonic acid is complete.
Moreover di↵erent ring precursor molecules were successfully synthesized. It has been found that the primary amine intermediate 17 is quite labile, resulting in only moderate yields for the final synthesis of ring precursor 18. Methoxy functioanlized ring precursor 41 has also been successfully synthesized via a pathway of hydroxylation, oxidation, grignard reaction and reduction, see scheme2.11.
It was also attempted to start o↵ the synthesis of this material with a double Du↵ reaction, however it was discovered that in order to force the reaction to completion harsh conditions are required. Mono substituted half product could not be completely suppressed even under the harshest conditions implemented in this study (see table2.4) and the byproduct could only be removed via column chromatography. The substrate scope and reaction conditions of the double Du↵ reaction were explored, reaching a number of conclusions. First of all the para position relative to the phenolic OH needs
to be blocked by a stable substituent in order to prevent formation of an aldehyde on the 4 position. It was discovered that iodine is not a satisfactory substituent because it is easily dissociated from the starting material under the acidic conditions of the Du↵ reaction. Unfortunately the unexpected findings were not investigated further in terms of mechanism, however dissociation of iodine was clearly observed. Once the reaction proceeds once and one aldehyde substituent has been installed on the substrate it de-activates it for further reaction, causing a lot of mono-formylated half products under various conditions. Paracetamol and paramethoxyphenol have shown that increasing the equivalents of hexamethylenetetramine, the reaction temperature and the reaction time can push the reaction towards the desired product, however still not to full completion. In conclusion, the Du↵ reaction works, ideally for products wherein mono-formylation is desired. One can force the reaction to proceed a second time towards a double Du↵ product, however the conditions for this are harsh.
3.2
Amine functionalized [2]rotaxane
In this part the first research question can be answered:
Can an amine functionalized ring be incorporated the capping mechanism for the covalent synthesis of [2]rotaxanes?
Regrettably the conclusion has to be drawn that following the current synthetic pathways a [2]rotaxane with NHBoc functional groups on the ring can not be synthesized. The NHBoc on the phenolic part of the ring fragment is not stable under hydrolysis conditions causing the ring to decompose in the last step. A solution could be to use another amine, or to methylate this substrate to a tertiary amine, possibly in an earlier stage. By this discovery the previous hypothesis of the ring closing on the wrong side of the rotaxane was disproven. The ring is lost on the grounds of decomposition, which is actually a good sign. If di↵erent side groups on the ring would cause the ring to close outside of the thread it could turn out arduous to make a diversified library of functionalized rings on the rotaxane. Yet it can for now still be claimed that the ring never closes outside of the thread or the template, demonstrating the strength of the terephatlic acid core template that is being used.
3.3
Methoxy functionalized [2]rotaxane
The next question that was answered is:
Can the ring in a [2]rotaxane, synthesized via covalent template assisted capping methods, be equipped with other functional groups?
A more joyous conclusion can be drawn from the research into the methoxy functional-ized ring on a [2]rotaxane, which was successfully synthesfunctional-ized via a capping mechanism. Low solubility of some intermediates cause the overall yield of the total synthesis to be only 13% (based on the terephtalic acid template). Nevertheless it is a very versatile synthetic route with many possibilities to research the scope of this type of rotaxanes. Di↵erent substituents on the ring, on the thread or on the stoppers can be installed. Higher order rotaxanes can be synthesized and a whole world toward molecular machines and switches is now opened.
Chapter 4
Outlook
Because of the research presented in this thesis, there is now more knowledge on the scope of the synthesis towards functionalized rotaxanes. This is not the end, but just the beginning of exploration of the potential that this system has to o↵er. The final goal of having a fluorescence based azobenzene switchable rotaxane was not reached, but there are now more tools and knowledge towards this goal. Investigation into the type of switches and incorporation into the rotaxane system needs to be done. The energy barriers of the switches can be tuned using di↵erent substituents, so calculations and experiments are necessary to investigate what will happen if these switches are build into the current rotaxane system. The same applies to the chromophores that are envisioned on the ring of the rotaxane. If this will become a BODIPY, its absorption and emmision wavelength can also be tuned via attachment of various side groups. Even if all this has been figured out, it needs to be determined in what stage of the synthesis the functional groups can be installed, so they do not interfere or decompose in later steps. Some suggestions have been provided in this thesis, however no experimental findings support the suggestions. In the current system there is no directionality, so even if it would be possible to trap the ring on one side of the thread there is no way to di↵erentiate between left or right. Di↵erent stoppers on either side of the rotaxane could be installed to induce recognition sites in the thread, however it would also be really innovative to have a directional ring and like so manipulate the movement of the ring over the thread. Hereby the doorway towards chirality in these rotaxanes could also be opened.
The final two questions that were set in the beginning of the project could not be an-swered. So to answer the following two questions a lot of further research needs to be executed:
Can recognition sites, in the form of a photoisomerizable group such as azobenzene moi-eties, be integrated in our [2]rotaxane system?
Can our current rotaxane system be functionalized in such a way that it can function as a fluorescence switchable molecular machine?
Further research will start o↵ by examining the reactivity of the methoxy group in the [2]rotaxane that has been synthesized in this research. Removal of the methoxy group to liberate the phenolic OH is necessary to achieve further functionalization. Then it should be investigated if a BODIPY group can be attached on this position on the ring and what the best conditions are to achieve this. Also the stage in the synthesis for this transformation needs to be determined, so before or after capping, before or after hydrolysis; this is all unknown. Another branch of research needs to be opened towards the switchable azobenzene groups in the thread and if they can be incorporated in the capping mechanism which is currently used. So still a lot of work, but thanks to this research we are now one small step closer to (hopefully) creating efficient molecular machines.
Acknowledgements
There are many people to thank. First of all my daily supervisor Martin Wanner, it was an honor to have such a great and experienced chemist as a supervisor. I learned a lot from him, I can truly say he knows everything when it comes to synthesis and practical work in the lab. From the little tricks that make an experiment run smoother to the reactivity of compounds and what to do if something goes wrong. But not only work-related, he also taught me to play cards which has been a nice break in the middle of the day. I would also like to thank him for his patience with me, taking the time to explain something to me if I did not understand it, taking the time to look at my reactions and spectra to figure out together what went wrong or right, but also having patience with me when I got stressed if something went ’di↵erently than planned’. Even after his retirement he was still willing to help me until even reviewing my thesis. I could not have wished for a better supervisor, I learned a lot from him, but most of all we also had fun! The second person that highly contributed to my research is prof. dr. Jan van Maarseveen. He o↵ered me a place in the group and has supported me throughout my whole internship. His door was always open for questions or any kind of help. Also the discussions in the bi-weekly meetings have been very helpful in keeping my research on track. Because of his intelligence and enthusiasm it has truly been a pleasure to do my internship in his group. He is very involved with the students in the group, which really helps to keep a motivated vibe in the lab, so thank you for that. I would also like to thank my second reviewer, prof. dr. Fred Brouwer. I asked him to be my second supervisor because of his in depth knowledge on the switches part of the molecular machine I attempted to synthesize. Unfortunately the research has not come to the point of working that out together, nevertheless thank you very much for taking the time to grade my thesis and presentation and being the second reviewer of this research. I feel I also was blessed with a second daily supervisor: Simone Pilon. On the moments Martin was out of the office he has been a great help and substitute supervisor. Our research lining up for a bit in the last couple of months of my internship has been very convenient for me, because he had great tips on my reactions. Even though the lab was busy with many students under his supervision he still took the time to also answer my questions and provide help when necessary which I highly appreciate! Thank you for all your help, I truly respect your hard work. Though only meeting him once, I would also like to thank dr. Luuk Streemers. The research he has done has been the whole foundation of my thesis and his PhD-thesis and labjounals have been my daily handbooks. He did such inspiring work and thank you for attending my presentation. His work has inspired a lot of other students as well, such as Steven Frolke, Tessa Roell, Mees Trouwborst, Inez Koornneef and Indigo Bekaert, all of whose thesis I have read (to some extent) and a lot of their work has been used in this thesis. So thank you to all the
previous students working on this project for laying so much ground work and basically making my life much easier. Especially Indigo Bekaert, her thesis, lab journal and even starting materials were used for the whole part on amine functionalized rotaxane. And as she was still doing an internship in the same building we also had some constructive conversations on the topic, so it was very nice continuing on her research. Besides all these people, the people who made my internship really fun were all the other students and employees on the lab. When I first started, the fumehood next to me was occupied by a rotaxane-building-machine: Milo Cornelissen. I don’t know how many exactly, but he assembled a large amount of rotaxanes in a small amount of time and was a great help because he knows all there is to synthesizing rotaxanes and was willing to share all his tips and tricks with me. Besides being smart, he is also really fun to work with and quickly became my go-to colleague for a quick co↵ee break and discussing everything from chemistry to politics. Other students working on rotaxanes have also worked alongside me during my stay in the group and have all been great colleagues which made the lab enjoyable and have been nice to exchange knowledge with; Kevin van Rijn, Chim Bijl, Sebastiaan Jager and Niels van Duijnen. Being all students in the lab figuring everything out we helped each other. Our lunch group was always a relaxing company to break the day and I feel over the course of my months in the group we all became friends and also solved many chemical challenges in our break. Wen-Liang Jia, thank you for translating chinese patents for me and helping me out finding my way around in ”the other lab”. Thomas Morsch, Klaas Visscher, Matthew Sultan and Bono van IJzendoorn, thank you for the fun times in the lab and also for relaxation after work, which is also very important. Even though we all did our internship in the same group on very di↵erent subjects you can still help each other out. Bas de Jong and Nick Westerveld are to be thanked for keeping the lab running. Special thanks to our ”klaverjas”-group, usually consisting of Hans Bieraugel, Martin Wanner and Dorette Tromp. When I started I did not even know how to play the game and now I am hooked. Spending time with this experienced group of kind chemists has been fun and educative! Last but not least in this long list of acknowledgements are the technicians who measured mass samples for me, Ed Zuidinga and Dorette Tromp. All the other people in group, in the labs and offices who have not been specifically mentioned are not less important. Each and every one of them contributed to my enjoyable stay in the research group, so thank you all SOC-members!
Appendix A
Supplementary Information
A.1
General methods and materials
All of the reactions were performed without taking any special precautions such as drying or N2/Argon atmosphere, unless stated in the procedure. The dry solvents CH2Cl2 and CH3CN were distilled using CaH2as the drying agent. THF was dried by ditilling it with sodium as the drying agent. All of the dry solvents are stored under N2atmosphere. The
dry solvents DMF and DMSO are obtained from Sigma Aldrich and stored under N2
atmosphere on 4˚A molecular sieves. Reagents were purchased (purity generally ¿98%) from Sigma Aldrich and Fluorochem and used as received. Grubbs 2nd generation catalyst was purchased from AK Scientific. Reactions were followed using thin layer chromatography (TLC) on 0.25 mm E. Merck silica gel plates (60F-254). SilaFlash P60 (particle size 40-63 µm) was used for flash column chromatography. NMR spectra were recorded on Bruker DRX-300, 400 and 500 MHz instruments and calibrated on residual undeuterated solvent signals. The 1H-NMR multiplicities are abbreviated as: s = singlet, d = doublet, t = triplet, q = quartet, quint = quintet, m = multiplet. High resolution mass spectra (HRMS) were recorded on a Mass spectra were collected on an AccuTOF GC v 4g, JMS-T100GCV Mass spectrometer (JEOL, Japan). FD/FI probe equipped with FD Emitter, Carbotec or Linden (Germany), FD 10 m. The current rate is 51.2 mA/min over 1.2 min machine using field desorption (FD) as ionization method. For compounds 17, 18, 43, 44, 45, 46 and 47 procedures from Bekaert, I. A. Covalent synthesis of a ring-functionalized [2]rotaxane, minor research report Master Chemistry Molecular Sciences, prof. dr. Jan van Maarseveen, Synthetic Organic Chemistry van t Ho↵ Institute for Molecular Sciences, 2018 were followed. For compounds 5, 6 and 7, procedures from Frolke S. A highly modular synthesis of [2]- and [3]rotaxanes, Bachelor thesis scheikunde, prof. dr. Jan van Maarseveen, Synthetic Organic Chemistry van t
Ho↵ Institute for Molecular Sciences, August 2018 were followed. For compounds 11, 12, 13 14 procedures from Steemers, L.; Wanner, M. J.; Ehlers, A. W.; Hiemstra, H.; Maarseveen, J. H. van. Org. Lett. 2017, 19, 2342 were used.
Compound 5
Compound 1 (10 g, 44 mmol, 1 eq.) and allylbromide (9.6 mL, 111 mmol, 2.5 eq.) were dissolved in 90 mL DMF and K2CO3 (15.3 g, 111 mmol, 2.5 eq.) was added. This was stirred overnight at 50°C. An extraction was done with 2x 300 mL ehtyl acetate. Combined organic fractions were washed with water (2x 200 mL) and brine
(1x 200 mL), dried over MgSO4 and concentrated in vacuo. NMR
matched literature, so it was continued without further purification [56].
Compound 6
Compound 5 was dissolved in 200 mL of a mixture of
THF/H2O/Methanol (2:1:1). KOH (9.9 g, 176 mmol, 4 eq.) was
added and stirred for 20 minutes before all compounds dissolved. This was stirred for 3 hours at room temperature whereafter TLC showed full conversion. 17 mL conc. HCl solution was added and this was stirred for 10 more minutes. The mixture was dilluted with 100 mL water and extracted with ethyl acetate (2x 200 mL). The combined organic fractions were washed with brine (1x 200 mL), dried over MgSO4 and concentrated in vacuo. The product was dried on the oil pump overnight to remove residual solvent to yield 6 (9.49 g, 78% over two steps). NMR matched literature, so there was continued without further purification [56].
Compound 7
Pentaflourophenol (7.9 g, 43.13 mmol, 3 eq.), compound 6 (4.0 g, 14.4 mmol, 1 eq.), HBTU (16.3 g, 43.13 mmol, 3 eq.) and DIPEA (15.4 mL, 86.25 mmol, 6 eq.) were dissolved in dry THF. This was stirred at room temperature under N2 atmosphere over the weekend. The solvent was removed via rotary evaporation. The product was partitioned between 150 mL ethyl acetate and 150 mL water and the water layer
was extracted with ethyl acetate (3x 200 mL). Combined
organic fractions were washed with 150 mL 1M HCl, 150 mL concentrated NaHCO3,
150 mL brine and dried over MgSO4. After concentration in vacuo the product showed multiple spots on TLC so it was triturated with cold cold ethyl acetate to collect 7.1 g 7 (81% yield). NMR matched literature [56].
Compound 11
1-Bromo-4-tert-butylbenzene (10.4 mL, 60 mmol, 3 eq.) was dis-solved in 120 mL dry THF and cooled to 0°C. n-BuLi 1.6M in hex-anes (37.5 mL, 60 mmol, 3eq.) was added slowly and the mixture was stirred for one hour. Dimethyl carbonate (1.7 mL, 20 mmol, 1 eq.) was added dropwise and the reaction mixture was allowed to go to room temperature and was stirred overnight. After concentration in vacuo the residue was partitioned between 250 mL ethyl acetate and 120 mL 1M HCl. The water layer was washed with 60 mL ethyl acetate and the combined organic fractions were washed with 120 mL concentrated NaHCO3 solution, water (1x 120 mL) and brine (120x mL), then dried over MgSO4 and concentrated in vacuo. The residue was triturated with 80 mL cold petroleum ether to yield 11 (6.4 g, 15 mmol, 75% yield). NMR matched literature [35].
Compound 12
11 (6.5 g, 15 mmol, 1 eq.) and malonic acid (16 g, 150 mmol, 10 eq.) were heated to 180 °C in a 250 mL flask. The mixture started to evolve CO2 when the two white solids started to melt to form a dark yellow/orange viscous liquid. This was stirred for 3 hours until no more gas formation was observed and the mixture was dry again. The heating was removed and the solid was stirred until a dry slightly yellow solid was formed. This was air dried overnight and then dissolved in 150 mL ethyl acetate, washed with water (2x 50 mL) and brine (1x 50 mL), dried over MgSO4 and concentrated in vacuo. The residue was triturated with cold methanol to yield 12 as a white solid (6.4 g, 13.6 mmol, 90% yield). NMR matched literature, however showed starting material peaks [35].
Compound 13
12 (5.9 g, 13.3 mmol, 1 eq.) was dissolved in 100 mL dry TFH under N2atmosphere and was cooled to 0°C. BH3· SMe2(3.1 mL, 33 mmol, 2.5 eq.) was added dropwise and the mixture was stirred overnight at room temperature. The reaction was quenched by slowly adding 10 mL water and stirring this 30 minutes. The solvent was removed in vacuo and the residue was partitioned between 120 mL ethyl acetate and 75 mL water. The water layer was washed with 30 mL ethyl acetate and the combined organic layers were washed with 75 mL brine and dried over MgSO4. Concentration in vacuo yields a white foam which was purified using column chromatography (PE/DCM 1:1 to pure DCM). The first fraction was compound 11 and the second fraction was product 13 (3.1 g, 6.8 mmol, 51% yield). NMR matched literature [35].
Compound 14
Compound 13 (3.1 g, 6.8 mmol, 1 eq.) was dissolved 70 mL dry THF
under N2 atmosphere. The mixture was cooled to 0 °C and PPh3
(2.0 g, 7.5 mmol, 1.1 eq.) and DIAD (1.5 mL, 7.5 mmol, 1.1 eq.) were added. This was stirred for 5 minutes before adding DPPA (1.6 mL, 7.5 mmol, 1.1 eq.) dropwise. The reaction was stirred overnight at room temperature. The mixture was concentrated in vacuo and purified via column chromatography (PE/DCM 7:1) to yield 14 (2.5 g, 5.1 mmol, 75% yield). NMR matched literature [35].
Compound 17
Under N2 atmosphere, 16 (0.5 g, 1.5 mmol; 1.0 eq) and
SnCl2 2 H2O (1.7 g, 7.5 mmol; 5.0 eq) were dissolved in 15 mL EtOH and refluxed for 23 h. After cooling down the re-action mixture was made basic (pH=8) with Na2CO3. After filtration over celite to remove the tin salts the mixture was extracted with ethyl acetate (3x 50 mL) and the combined organic layers were washed with brine (2x 50 mL), dried over MgSO4 and concentrated in vacuo to yield a 17 (0.4 g, quantitative). NMR matched literature [57].
Compound 18
Compound 17 (0.6 g, 1.9 mmol, 1.0 eq) was dissolved in 20 mL dry DCM under N2. Boc2O (0.5 mg, 2.3 mmol, 1.2 eq) was added and the reaction was stirred at 40°C overnight. After evaporation in vacuo the product was purified via column chromatography (PE/EtOAc 30:1 to 20:1 to 10:1) as a slightly pink solid (0.4 g, 1.1 mmol, 56%). NMR matched literature [57].
Compounds 27-30
This is a general procedure for the formation of products 27 through 30 from phenol. Various conditions have been applied in terms of temperature, equivalents of HMTA and reaction time, see table2.1. Phenol (0.94 g, 10 mmol) and HMTA (various amounts) were dis-solved in 30 mL TFA. This was stirred at elevated temperature to yield a mixture of products. Workup consists of stirring the mix-ture with water for 1 hour, extraction with DCM (3x 30 mL) and concentration in vacuo. Compound 27(R1, R2 and R3 = aldehydes): 1H NMR (300 MHz, Chloroform-d) 12.19 (s, 1H), 10.35 (s, 2H), 10.03 (s, 1H), 8.54 (s, 2H). Compound 29(R1 and R2 = aldehydes, R3 = H): 1H NMR (300 MHz, Chloroform-d) 11.58 (s, 1H), 10.00 (d, J = 19.0 Hz, 2H), 8.21 8.05 (m, 2H), 7.17 (d, J = 8.7 Hz, 1H). Compound 30 (R1 = aldehyde, R2 and R3 = H): 1H NMR (300 MHz, Chloroform-d) 11.05 (s, 1H), 9.93 (s, 1H), 7.63 7.49 (m, 2H), 7.11 6.97 (m, 2H).
Compounds 31-34
This is a general procedure for the formation of products 31 through 34 from 4-iodo-phenol. Various conditions have been applied in terms of temperature, equivalents of HMTA and reaction time, see
table 2.2. 4-Iodophenol (0.5 g, 2.3 mmol) and HMTA (various
amounts) were dissolved in 15 mL TFA. This was stirred at ele-vated temperature to yield a mixture of products. Workup consists of stirring for one hour in 15 mL 4M HCl at 0°C. Then a filtration, residue contains the product(s). Compound 33: (R1 and R2 = aldehyde, R3 = I): 1H NMR (300 MHz, DMSO-d6) 10.08 (s, 1H), 9.87 (d, J = 2.0 Hz, 2H), 8.53 (s, 1H), 8.34 (d, J = 2.1 Hz, 1H). Compound 34 (R1= aldehyde, R2 and R3 = I): 1H NMR (300
MHz, Chloroform-d) 11.77 (s, 1H), 9.73 (s, 1H), 8.28 (d, J = 2.1 Hz, 1H), 7.87 (d, J = 2.1 Hz, 1H).
Compounds 35-36
This is a general procedure for the formation of products 35 and 36 from paracetamol. Various conditions have been applied in terms of temperature, equivalents of HMTA and reaction time, see table2.3. Paracetamol (0.2 g, 1.3 mmol) and HMTA (various amounts) were dissolved in 10 mL TFA. This was stirred at elevated temperature to yield a mixture of products. Workup consists of stirring the mixture with 2M HCl for 1 hour, extraction with DCM (3x 25 mL), washing the combined organic layers with 2M HCl (2x 25 mL) and brine (1x 25 mL), drying over
MgSO4 and concentration in vacuo. Compound 35 (R1= aldehyde, R2 = H): 1H NMR
(300 MHz, Chloroform-d) 10.87 (s, 1H), 9.86 (s, 1H), 8.00 (d, J = 2.6 Hz, 1H), 7.62 (s, 1H), 7.41 (dd, J = 8.9, 2.6 Hz, 1H), 6.94 (d, J = 8.9 Hz, 1H), 2.20 (s, 3H). Compound 36 (R1 and R2 = aldehydes): not isolated, however the characteristic phenolic OH peak has shifted to 11.87 (s, 1H), and the aldehyde peak has doubled in integral and also shifted upfield to 10.26 (s, 2H).
Compound 37-38
This is a general procedure for the formation of products 37 and 38 from paramethoxyphenol. Various conditions have been applied in terms of temperature, equivalents of HMTA and reaction time, see table2.4. Paramethoxyphenol (0.3 g, 2.4 mmol) and HMTA (various amounts) were dissolved in 5 mL TFA. This was stirred at elevated temperature to yield a mixture of products. Workup consists of stirring the mixture with 12 mL 4M HCl for 10 minutes, extraction with DCM (3x 25 mL), washing the combined organic layers with 4M HCl (1x 25 mL), water (1x 25 mL) and brine (1x 25 mL), drying over MgSO4and concentration in vacuo. Compound 37 (R1 and R2 = aldehydes): 1H NMR (500 MHz, Chloroform-d) 11.15 (s, 1H), 10.25 (s, 2H), 7.54 (s, 2H), 3.89 (s, 3H). Compound 38 (R1= aldehyde, R2 = H): 1H NMR (500 MHz, Chloroform-d) 10.67 (s, 1H), 9.89 (s, 1H), 7.17 (dd, J = 9.0, 3.1 Hz, 1H), 7.03 (d, J = 3.1 Hz, 1H), 6.96 (d, J = 9.0 Hz, 1H), 3.85 (s, 3H).
