Chapter 1: Catenane

The thesis initially focussed on the synthesis of the catenanes (S,S)-72 and (S)-73. For this purpose, the phosphoric acid macrocycle (S)-67 was synthesised in the first steps. This was done with an overall yield of 3% over 7 steps. The BINOL amine (S)-142 was also synthesised from (S)-BINOL. An established synthesis route was used first. However, this only produced a yield of 1% over 8 steps. In optimisation trials, the regulations were adapted and an increase in yield to 3% over 8 steps was achieved. The next step was the coupling of the ethylene glycol linkers. Various approaches were trialled here. On the one hand, the coupling of a boronic acid ester 91, whereby the coupling would take place in one step, and on the other hand, the coupling of 4-hydroxyphenyl boronic acid and the subsequent coupling of the ethylene glycol linkers. The one-step synthesis route provided yields of 5%, whereas the two-step synthesis route enabled a yield of 45%.

The dibenzylammonium thread 119 was produced in 4 steps with a yield of 40%. With the precursors in hand, the catenane syntheses were then started. The established method of ring closure metatesis was used for this purpose. Although the crude spectrum looked promising, it was realised that no catenane could be isolated. It was then considered to first form the amine macrocycle and then to form the catenane in a second ring closure of the phosphoric acid precursor. As the ring closure of the phosphoric acid was also an unknown reaction, tests were carried out first. For this purpose, the open phosphoric acid was reacted with Grubbs II and the reaction was observed using HPLC. It turned out that the simple ring closure of phosphoric acid (S)-124 produces by-products in addition to the desired macrocycle. When an amine (dibenzylamine) was added, it was found that only the macrocycle was formed. This suggests that there is a good interaction between the phosphoric acid and the amine, which suppresses side reactions.

The synthesis of the catenanes proved to be more demanding than originally assumed and posed a non-trivial challenge. Despite intensive efforts, numerous synthetic approaches and strategies for solving problems that occurred, it was ultimately determined that due to the limited time resources of this work, the desired goal of catenane synthesis could not be achieved. Although the detailed synthesis could not be realised, the challenges encountered during the research process, the approaches taken and the knowledge gained so far are discussed in this work. The documented information not only serves to illustrate the difficulties involved in the synthesis of catenanes, but also provides a valuable basis for future research in this area. To conclude, it can be said that the ring closure of the amine precursor with the free amine does not work in this way. The ring closure must be carried out with the Boc-protected derivative (S)-110 and only then can the deprotection be carried out in order to obtain the amine macrocycle. The amine macrocycle must be produced first. This suggests that the rotaxane synthesis in a final step should be promising. For the continuation of this project, the first step would be to optimize the reaction steps of the amine macrocycle and the final synthesis. As the Boc-protected macrocycle (S)-123 has already been synthesised, the only thing missing here is the deprotection and the production of the PF₆ salt. Once the catenane has been successfully synthesised, the properties can then be investigated. As the results of the corresponding rotaxane in the Michael addition were not promising, other areas of application should be sought here. However, the influence of the interlocking between catenane and rotaxane could also be investigated here.

For the second catenane with the benzylamine macrocycle, a little more work would have to be invested to continue. Firstly, amine strand 118 would have to be protected and then the ring closure would also have to be tested here. This could prove to be more difficult, as the overall more flexible structure of 118 compared to azepine (S)-81 could favour oligomerization over an intramolecular reaction. Once you have this macrocycle in hand, you can proceed as with the other catenane.

Chapter 2: Rotaxane

The second part of the work focused on the synthesis of various rotaxanes. The first rotaxane (S,S,R,S)-74 to be synthesised consists of a prolinol thread and a BINOL-based crown ether (S,S)-129. The idea was that a chiral rotaxane for catalysis could be formed from simple and quickly accessible building blocks. The individual building blocks and the associated syntheses are known from the literature and could be realised quickly. As a result, the macrocycle was obtained in 3 steps with a yield of 20%. The amine building block was also synthesised in 3 steps with a yield of 47%. This was followed by the synthesis of rotaxane via a copper-catalysed “Click” reaction. After the first test, only the macrocycle (S,S)-129 could be recovered. In a second test, the amount of amine (R,S)-133 and stopper 126 was increased. This time, thread (R,S)-134 was also obtained in addition to the macrocycle. Unfortunately, rotaxane (S,S,R,S)-48 could not be synthesised. To optimise this synthesis, it would be possible to increase the interaction between thread and macrocycle. If the interaction improves, a pre-orientation of the pseudorotaxan is more likely. This can be done by introducing groups that preferentially form hydrogen bonds like benzylic CH₂ groups.

Rotaxane (S,S)-75 was synthesised as the next part of this chapter. This is based on the known macrocycle with isopropyl groups (S)-68 from preliminary work by Niemeyer and a BINOL-based amine axis (S)-145. The macrocycle was synthesised in 8 steps and gave a yield of 7%. The stopper was obtained in 4 steps with a yield of 45%. The amine axis is based on the same building block as the previously described catenane. For this purpose, the Boc-supported amine (S)-80 was converted to the hexafluorophosphate salt (S)-144 in 4 further steps with an overall yield of 50%. First, the thread (S)-145 was successfully synthesised (yield: 62%), which was to serve as a comparison for the analyses. The rotaxane synthesis was then started. Rotaxane (S,S)-76 was successfully obtained with a yield of 28%.

Three further rotaxanes were synthesised, which are based on a proline moiety and the same macrocycle (S)-67 or (S)-68 or (R)-68. For the axis, (l)-prolinol was started and obtained in 3 steps with an overall yield of 60%. First, the thread (R,S)-146 was synthesised, which was obtained with a yield of 46%. As the other building blocks were already available, the rotaxane synthesis could be started directly. All three rotaxanes were obtained with a yield of (S,R,S)-77: 17%, (S,R,S)-78: 37% and (R,R,S)-78: 19%.

With the four rotaxanes in hand, catalysis experiments were then carried out. The Michael addition was used to compare the results with the results from Niemeyer's previous work. Firstly, cinnamaldehyde and secondly a naphthyl derivative were reacted with malonate. In the catalysis experiments, it was found that the rotaxane based on BINOL (S,S)-76 did not produce any conversion. The enantiomeric excess when using the rotaxane (S,R,S)-77 was also rather low (-6% ee), which meant that the two rotaxanes were neglected for further catalysis experiments. The rotaxane catalysts were able to convert both substrates 152a/b into the corresponding products 147a/b. The rotaxanes are significantly more active than either the free prolinol-thread or the non-interlocked mixture of thread and either enantiomer of the macrocycle with a conversion of around 71 - 84% after 7 days. The thread alone had a conversion of 12 - 34% after 7 days and the non-interlocked mixtures of the subcomponents had a conversion of 7 - 30%. If the rotaxanes are compared with their non-linked components, the following statements can be made: Firstly, the (R,S)-prolinol thread alone only leads to a low enantioinduction , for 152a 14% and for 152b 11% ee. When the two individual components are combined, the enantioselectivities for the non-interlocked mixture remain largely unchanged, 21% ee for 152a and 15% ee for 152b. If one compares the results with those obtained by Niemeyer with the rotaxane (S)-71, it can be seen that in both cases the interlocked nature of the catalyst leads to a cooperative behaviour of the lithium phosphate and the amine group. This results in an increase in the reaction rates. However, the inclusion of the chiral prolinol thread did not lead to an increase in stereoselectivity compared to the rotaxane with the achiral thread. On the other hand, the results clearly show matched/mismatched effects for the diastereomeric rotaxanes (S,R,S)-78 and (R,R,S)-78 for both 152a (8%/29 % ee) and the naphthyl derivative 152b (30 %/0 % ee).

If the project is continued, it may be worth considering testing the rotaxanes in other catalyses to see which reaction these catalysts are best suited for. In addition, other chiral amine threads could be prepared and their rotaxanes tested for matched and mismatched activities.