Attempted synthesis of a meta-metalated calix[4]arene

An evidence for the formation of a rare meta-metalated inherently chiral calix[4]arene is described. Our strategy involved using a mesoionic carbene to direct C–H activation, but proved to form an unexpectedly unstable intermediate that was identified through high-resolution mass spectrometry. On route to our target, a new optimized method to mononitrocalix[4]arenes was developed, including optimized and high yielding transformations to azide and 1,2,3-triazole derivatives which may have application in other areas of research.

The characterisation data collected for this compound compared well to literature data.  A 2-neck round bottom flask (100 mL) was charged with ethanol (14 mL), mononitrocalix [4]arene 9 (455 mg, 0.713 mmol) and 10% palladium on carbon (114 mg, 0.107 mmol, 0.15 equiv) sequentially. The mixture was heated under reflux and hydrazine hydrate (181 µL, 3.70 mmol, 5.2 equiv) was added. The reaction mixture was heated under reflux for 3 hours after which it was cooled to rt, filtered through Celite, and washed with DCM (20 mL). The solvent was removed in vacuo and the product dried under high vacuum, affording monoaminocalix [4]arene 10 as a white solid (431 mg, 0.71 mmol, 99% yield).
The characterisation data collected for this compound compared well to literature data.  95-97% H2SO4 (105 µL, 1.97 mmol, 3.0 equiv) was added to a solution of NaNO2 (182 mg, 2.63 mmol, 4.0 equiv) in H2O (1.0 mL) at 0 °C. Bubbling of the solution was visible and a brown gas formed immediately upon addition of the acid. The solution of NaNO2 and H2SO4 was added dropwise to a solution of monoaminocalix [4]arene 10 (400 mg, 0.658 mmol) in THF (2.0 mL) and MeCN (2.0 mL) at 0 °C, over the space of 5 minutes. The colour of the solution changed to a dark orange after 15 minutes, after which NaN3 (129 mg, 1.97 mmol, 3.0 equiv) in H2O (0.5 mL) was added dropwise over 5 minutes. The solution immediately became S15 a murky peach colour, with bubbles of N2 evolving. The reaction mixture was stirred for 2 hours while warming to room temperature, after which H2O (10 mL) was added to quench the reaction. The product was extracted with EtOAc (3 × 20 mL) and the combined organic layers dried over MgSO4. The solvent was removed in vacuo, yielding monoazidocalix [4]arene 13 as an orange solid (409 mg, 0.645 mmol, 98% yield).   (7). Various attempts were made to assign the unknown signals, but the predicted formulas made no sense based on possible structures, but probably represent unusual adducts.

Rf
The same equivalents of all reagents were added and the mixture was heated to 85 °C. Even at 85 °C the solubility of the starting material was poor and monobromocalix [4]arene was visible as a white solid floating around the reaction flask. The reaction was left to stir overnight, but TLC the next day indicated that no product had formed. It was postulated that there was too much water present in the reaction, which prevented the starting material from dissolving effectively. The hydrophobic nature of the starting material meant that the logical response was to repeat the reaction using less water. This was done by dissolving 200 mg of monobromocalix [4]arene 6 in 8 mL of acetone and 0.5 mL of distilled water, which resulted in a solvent ratio of 16:1 (entry 1, Table S1). The starting material dissolved instantly, which showed that a more dilute solution with less water was a step in the right direction in terms of solubility. The same equivalents of all reagents were added and the reaction mixture was degassed before being heated. At around 55 °C, the solution was dark turquoise in colour and became completely clear when heated to 85 °C. TLC analysis after 4 hours using 2% EtOAc/PET showed the formation of a single new spot which was slightly more polar than the starting material spot. A 10% solution of PPh3 staining technique revealed that the lower spot was an azide. TLC indicated a substantial amount of remaining starting material, so the reaction was left to stir overnight. An almost identical result was obtained from TLC the next morning, and the reaction had unfortunately not gone to completion. The reaction was stopped after 24 hours regardless, by addition of 1 M HCl to the dark green solution.
Purification via silica gel flash column chromatography turned out to be significantly more challenging than expected due to the similarity in Rf values of the starting material and product. Multiple attempts using different solvent systems all produced co-eluted fractions, which contained both starting material and product. FTIR analysis of the crude mixture produced an azide stretch at 2108 cm -1 which confirmed that the azide had formed, however, quantification had not yet been achieved. It was speculated that the reaction had been diluted too drastically when attempting to dissolve the starting material and this had inhibited reaction completion. The synthesis was repeated using the same 16:1 solvent ratio, reagent quantities and temperature as before (entry 2, Table S1). The concentration was increased by using 4.8 mL of acetone and 0.3 mL of distilled water however. TLC analysis after stirring under reflux for 24 hours at 85 °C indicated that a seemingly identical result to the previous reaction had been obtained.
The use of higher reagent quantities was next investigated. It was theorised that the hydrophobic cavity and sheer size of the calix [4]arene was preventing efficient contact of the reagents with the brominated active site. Table S1's third entry shows the conditions for the next attempted synthesis of monoazidocalix [4]arene 7. Monobromocalix[4]arene 6 was added to 8 mL of acetone and 0.5 mL of distilled water, followed by addition of all reagents with doubled equivalents: NaN3 (4 equiv), Na-ascorbate (0.1 equiv), CuI (0.2 equiv) and DMEDA (0.3 equiv). After the solvent was degassed, the mixture was stirred overnight under reflux at 85 °C. The reaction had not yet reached completion after 21 hours, so additional NaN3 (2 equiv) was added and the reaction mixture was stirred for a further 5 hours. Unfortunately, after a total reaction time of 26 hours with increased reagent quantities, it was found by TLC that there was still starting material remaining. A change in TLC eluent from 2% EtOAc/PET to 35% DCM/PET brought to light that the previously suspected single azide spot was in fact three spots with extremely similar Rf values. Two of the three spots were identified by co-spotting tetrapropoxycalix [4]arene 8 alongside the reaction mixture, and by using the previously described azide stain. Both the desired azide product and tetrapropoxycalix [4]arene 8 had formed, and were found to have essentially identical Rf values. It turned out that quenching of some kind had taken place during the reaction and formed the latter. The third compound, which was only marginally more polar than the two previously mentioned compounds, was not successfully identified. Solubility of NaN3 had become a potential concern due to its insolubility in acetone and it was therefore decided to add more water to the next reactions. A solution of monobromocalix [4]arene 6, NaN3 (2 equiv), Na-ascorbate (0.1 equiv), CuI (0.2 equiv) and DMEDA (0.3 equiv) was prepared in acetone (10 mL) and distilled water (1 mL) (entry 4, Table S1). The reaction mixture was stirred under reflux at 90 °C for 24 hours after which TLC confirmed that a similar result to the previous reaction had been obtained. Although the reaction had not reached completion, a positive outcome was that the starting material spot seemed less prominent than before. The reaction was repeated one more time in the acetone/H2O solvent system and stirred for 48 hours at 80 °C (entry 5, Table S1). This unfortunately did not lead to an improved result, so focus was shifted to a new solvent combination.
The Ullmann-type coupling reaction was next performed using a DMF/H2O solvent system. Monobromo-calix [4]arene was dissolved in DMF (8 mL) and distilled water (1 mL). NaN3 (2 equiv), Na-ascorbate (0.1 equiv), CuI (0.2 equiv) and DMEDA (0.3 equiv) were added to the flask and the mixture was heated to 100 °C after being degassed. The result after heating under reflux for 20 hours was a slightly cleaner reaction in which the unidentified compound from the previous reactions had not formed. The spot, which was thought to be the desired azide product, did not stain positively for an azide however, which did not bode well. It was considered that the copper catalyst was perhaps decomposing during the reaction. The synthesis was repeated using considerably higher reagent quantities in a more concentrated solution of DMF (5 mL) and distilled water (1 mL). NaN3 (4 equiv), Na-ascorbate (0.2 equiv), CuI (0.4 equiv) and DMEDA (0.3 equiv) were added to the flask containing monobromocalix [4]arene 6 and the mixture was heated to 100 °C after being degassed. 18 hours of stirring under reflux did not produce a positive result. Predominantly monobromocalix [4]arene starting material and a negligible amount of the target azide were visualised using TLC.

Attempted nitration with acetic acid
Examples of tetrapropoxy-calix [4]arene mononitration found in the literature are based on a procedure reported by Reinhoudt and co-workers. 5 The procedure involved stirring tetrapropoxycalix [4]arene and glacial acetic acid (42 equiv) in DCM. After addition of 65% nitric acid (15 equiv), the mixture was stirred at room temperature for 30 minutes before being quenched with water. The product was extracted with DCM, after which the organic layer was washed with Na2CO3 and water before being dried over MgSO4. Removal of the solvent in vacuo was followed by purification via silica gel flash column chromatography. From this, Reinhoudt and co-workers obtained a mixture of mononitrocalix [4]arene (30%) with traces of distal dinitro-calix [4]arene and proximal dinitro-calix [4]arene.
The first synthesis of mononitrocalix [4]arene 9 (entry 1, Table S2) was performed just as described by Reinhoudt and co-workers. After 30 minutes of stirring at room temperature (22 °C), the reaction was quenched by addition of water. Following the work-up, it was discovered that only starting material had been recovered. The reaction was repeated the next day using 70% nitric acid (entry 2, Table S2) and stirred at 23 °C. Slight darkening of the solution was observed, where before it had remained clear. The formation of a faint new spot near the baseline was observed from TLC and the reaction was quenched again after 30 minutes. A negligible amount of product had formed and it became apparent that a longer reaction time was necessary. Tetrapropoxycalix [4]arene was then stirred for 2 hours at room temperature (23 °C) in the presence of the two acids (entry 3, Table S2). After this time, TLC of the mixture showed the presence of starting material, a more prominent mononitrocalix [4]arene spot and a new spot which had formed near the baseline. After quenching with water, purification of the brown/orange crude mixture was achieved via silica gel flash column chromatography using an eluent gradient of 15% to 25% (DCM/PET). It was found that half of the tetrapropoxycalix [4]arene 8 starting material remained unreacted, and was obtained as the first compound. Mononitrocalix [4]arene 9 was obtained in the next fractions as a light yellow solid in only 21% yield. The column was then flushed to isolate the bottom spot, and a trace amount of what was presumed to be dinitrocalix [4]arene was obtained. The compound was not characterised, as the focus of this study was calix [4]arene mononitration.

S29
The next attempt at improving the yield of the nitration reaction involved stirring tetrapropoxycalix [4]arene 8 in DCM with both acids at room temperature (20 °C) while monitoring the progress of the reaction every hour (entry 4, Table  S2). After 7 hours of stirring, no product was visible and the solution remained clear. The reaction mixture was then left to stir overnight at room temperature. TLC the next morning of the pitch-black solution indicated that overnitration had occurred throughout the night. No starting material was visible and only a trace amount of mononitrocalix [4]arene 9 could be observed. Both species had been converted to over-nitrated products, which had formed a prominent spot on the baseline. The TLC eluent's polarity was increased from 20% to 60% (DCM/PET) to move the spot off the baseline. Three compounds became visible, of which two had extremely similar polarities. It was theorised that these were both proximal-and distal dinitrocalix [4]arene, based on the findings of Reinhoudt and coworkers. The more polar spot was suspected to be trinitrocalix [4]arene. The synthesis was then repeated (entry 5, Table S2) while monitoring reaction progress via TLC every 30 minutes. As there was a distinct lack of control over the reaction in terms of over-nitration, a standardised method of TLC was performed to try and quantify formation of products by UV visualisation. A constant volume of 100 μL was removed from the reaction mixture every 30 minutes for 2.5 hours. These volumes were then diluted with a constant volume of DCM (0.1 mL) and spotted twice on the silica TLC plate using the same spotter. The idea was to quench the reaction when it seemed like the rate of mononitrocalix [4]arene 9 formation had decreased and the rate of over-nitrated product formation had increased. It was a shot in the dark as there was no way of knowing the relationship between these two rates. It was also not possible to know whether over-nitration was occurring on the starting material, mononitrocalix [4]arene or both. After the first hour of stirring at 20 °C, the colour of the solution had become much darker but no product was visible from TLC. In the next hour, a faint product spot had formed. The spot had become noticeably darker within the next 30 minutes, but the formation of a by-product which was only slightly more polar than the desired compound was also noted. The reaction was stirred for a further 15 minutes before being quenched, which marked a total reaction time of 2 hours and 45 minutes. At this point, the by-product spot had become considerably darker and spots of overnitrated products had started to form on the baseline. Formation of a spot, which was extremely close to the product spot, meant that co-elution was a concern when purifying the crude mixture. For this reason, the volume of silica gel was increased when flash column chromatography was performed. Mononitrocalix [4]arene 9 was obtained as pure compound in 28% yield, with just more than a third of the starting material which remained unreacted. This was an improved result in terms of yield and quantity of unreacted starting material. The synthesis was then repeated in a similar fashion (entry 6, Table S2). TLC was only performed after 1.5 hours, as not much product had formed up until this point in the previous synthesis. The solution had become pitch-black in colour after 2 hours and 50 minutes of stirring, and the same by-products were visible from TLC. The reaction was quenched for fear of losing calix [4]arene 9 to over-nitrated products. It was noted that the room temperature had increased from 20 °C to 23 °C throughout the reaction. After purification via silica gel flash column chromatography it was discovered that less than a third of the starting material remained unreacted and mononitro-calix [4]arene 9 had formed in 44% yield. This was a clear improvement in terms of yield and unreacted starting material. It was theorised that a temperature increase of merely 3 °C throughout the reaction had led to a significant increase in yield, even though the reaction time was only 5 minutes longer.

S30
At this point it had become clear that darkening of the solution was an indication that nitration had started to occur. Stirring of the next reaction was started at a room temperature of 23 °C (entry 7, Table S2). The solution remained clear after 35 minutes of stirring, so the reaction was heated to 30 °C and stirred for a further 25 minutes. No colour change was observed, so the reaction was quenched with water and the starting material extracted with DCM. This result was compared to entry 4 of Table S2, of which the solution remained clear for 7 hours. Over-nitration had then occurred when the reaction was left to stir overnight. It seemed as if the nature of the nitration was in some way autocatalytic, and that the reaction would not proceed if the initial nitration did not occur. It was decided to repeat the reaction with the extracted starting material. If a colour change was not observed after 1 hour of stirring, the reaction would be quenched. The next two reactions (entries 8 and 9, Table S2) provided nothing but frustration. After stirring for 1 and 3 hours respectively, no colour change was observed for either reaction. The inconsistency of this ** It should be noted that 65% nitric acid was used for this reaction. The remaining entries all used 70% nitric acid.

S31
reaction would not suffice when generating mononitro-calix [4]arene 9 material for use in subsequent reactions. A new nitration method was therefore investigated.
Attempted formation of the ruthenacycle 13 Figure S33. Zoomed in section of the HRMS spectrum showing the predicted (bottom) and acquired (top) section that corresponds to the ruthenacycle.
S32 Figure S34. Stacked plot of NMR spectra showing the ruthenium dimer (top), reaction mixture (middle) and starting material (bottom). Important points are: 1: disappearance of the triazole methine indicating C-M bond formation; 2: shifts in the p-cymene signals consistent with complex formation; and 3: signals (*) that one might expect for diastereomers when p-cymene is part of a ruthenacycle (speculative, but suggests very low yield).