C–H activation: a Critical Evaluation of a Published Method and its Application Towards Inherently Chiral Calix[4]arenes

C–H activation offers an intriguing access into inherently chiral calix[4]arenes, but has been little explored in the literature. In this article, we report our investigation into a published C–H activation method that uses carbamates to direct a palladium catalyzed C–H activation and subsequent reaction with N-bromosuccinimide. However, we show that this report is unfortunately flawed on a number of points. An earlier reported study revealed the more likely SEAr mechanism of the bromination reaction, which did not involve palladium catalysis. We nevertheless employed the SEAr bromination in an attempt to form inherently chiral calix[4]arenes, using a chiral (+)-menthyl carbamate as a directing group. Unfortunately, although the reaction was high yielding, the diastereomers formed were inseparable and we were unable to quantify their ratio. Subsequent removal of the chiral (+)-menthyl carbamate, returned a small positive optical rotation, suggesting that at least a level of asymmetric induction was achieved in the bromination to afford a non-racemic product.

co-workers was itself correct (Scheme 1). We had some initial concerns about the reported method, as the paper appeared to have superficial errors that we found surprising. Some of these errors might have been due to topographical oversight, but did warrant further investigation. One concern was that the reported method should have been theoretically possible using just N-bromosuccinimide (NBS), without any need for the palladium catalyst. In the paper, the authors reported that the reaction failed using NBS in acetonitrile, but only worked when palladium acetate was added. The authors then reported that the reaction failed in dichloroethane (DCE) even with palladium acetate, but worked again when para-toluenesulfonic acid (PTSA) was added. The published table of results showed no example of an experiment that then excluded the palladium catalyst, but kept PTSA, i.e. a control experiment. However, the text did report that a control experiment had been performed and then referred to an incorrect entry on the table (hence a possible typographical error). For this reason, we decided to have a closer look at the reaction ourselves.
The model selected (carbamate 1) included a para-methoxy group, which served both as a model for a single functionalised aromatic ring on the calix [4]arene, and as a means for testing the directing ability of the carbamate vs. the methoxy group. Essentially, it was found that the role of the palladium in this experiment was greatly exaggerated (see Table 1), with the yield of brominated 2 only being marginally higher when it was included. In both cases, the carbamate was the sole director towards ortho-bromination. It therefore seems likely to us that Moghaddam and co-workers had somewhat overstated the importance of the palladium and its role in the reaction.
With this rather unsurprising result, we took a much closer look at the paper by Moghaddam and co-workers and noticed more problems. In the introduction, they made the main claim that: 'To the best of our knowledge, this is the first report of application of N-arylcarbamates as DG in C-halogen bond formation'. This statement cannot be proven false, since it is 'to the best of their knowledge', but it is nevertheless wrong. A quick search on Reaxsys reveals a different story: excluding papers reported after their own 2016 publication, 43 documents (including patents) report the use of NBS brominating an aryl ring ortho to a carbamate; 12 documents using NCS (chlorination) and 24 documents using NIS (iodination). Many further examples can also be found employing the respective molecular halogen reagent (e.g. chlorine, bromine and iodine). A minor selection of examples from the peer review literature are shown in Fig. 2 (refer to Supplementary Information). It is disheartening that the reviewers never noticed this, since this fact alone puts a completely different interpretation onto the results presented.
Secondly, Moghaddam and co-workers cite a 2014 paper by RESEARCH ARTICLE K.J. Visagie, L. Hodson and G.E. Arnott, 16 S. Afr. J. Chem., 2020, 73, 15-21, <https://journals.co.za/content/journal/chem/>.  DG(sic).' This statement is false, since Uhlig and Li very definitely reported on N-arylcarbamates being used as C-H activation directing groups. In fact, it is the entire focus of their paper, which is titled 'Aniline Carbamates: A Versatile and Removable Motif for Palladium-Catalyzed Directed C-H Activation'. In Uhlig and Li's paper, they also had a good look at the reaction mechanism and found that the aniline carbamate strongly favoured the promotion of electrophilic aromatic substitution, which aligns with the observations in the literature that NXS is itself capable of halogenating ortho to an N-arylcarbamate via a non-C-H activation pathway. Thirdly, closer inspection of Moghaddam and co-workers' proposed mechanism also reveals a number of problems. Firstly, an intermediate involving a deprotonated carbamate is proposed, which is unlikely since the reaction is under acidic conditions, and secondly, they make no account of the purpose of the p-toluene sulfonic acid which they point out is crucial; see ref. 8 : Scheme 5. A more realistic and plausible mechanism can be found in the paper by Uhlig and Li, which they cited. 16 Overall, it is our opinion that Moghaddam and co-workers submitted a paper that failed to acknowledge the correct state of the art, and in so doing failed to understand the more likely interpretation of their results. The work by Uhlig and Li makes it clear that C-H activation is almost certainly occurring, but in the case of halogenation, this may not be the main mechanistic pathway. The unfortunate thing is that this paper has been cited many times, including four reviews where its conclusions have been reported without question. [17][18][19][20]

Calix[4]arene Study
Whilst our starting point for the study proved to be somewhat spurious, the use of a chiral carbamate to potentially form inherently chiral calix [4]arenes was deemed to be worth pursuing. First an achiral model study was carried out, in order to check the chemistry, by reacting the known mono-aminocalix [4]arene 3 21,22 with methyl chloroformate and pyridine (Scheme 2). The carbamate product 4 was confirmed via 1 H NMR spectroscopy (singlet at δ3.69 ppm for the hydrogen atoms of the methoxy group), infrared (1727 cm -1 for the carbamate) 23 and HRMS (calculated for C 42 H 52 NO 6 [M+H] + : 666.3790; found 666.3782). With this material in hand, we attempted the bromination using both Moghaddam and co-workers' method, against one that excluded the palladium catalyst. Once again, the yields of the reaction were only marginally higher when the palladium catalyst was included, suggesting that the dominant mechanism was still the electrophilic bromination and not C-H activation (see Table 2). What was also worth noting was that the unfuctionalized aromatic rings were not brominated, even though they are in principle activated by the para-propoxy groups. This is again likely due to the greater directing effect of the carbamate as evidenced in the model study (see earlier). Successful synthesis of the ortho-brominated product 5 was primarily concluded from the HRMS (calculated for C 42 H 51 BrNO 6 [M+H] + : 744.2900; found 744.2872 with matching isotopic distribution), and 1 H NMR spectroscopy (whose aromatic region revealed the loss of one proton).
Having established that the carbamate was a suitable functional group for activating the calix [4]arene meta-position, we turned our attention to using a chiral carbamate to see whether diastereoselectivity could be achieved. For this we selected (1S)-(+)-menthyl chloroformate (it was available in our labs), and reacted it with mono-amine calix [4]arene 3 under the same conditions as before. The new chiral carbamate 6 was obtained in excellent yields between 90 and 98 % after work-up and column chromatography. The mono-menthyl carbamate calix [4]arene 6 was characterized by NMR spectroscopy, HRMS (calculated for C 51 H 68 NO 6 [M+H] + : 790.5047; found 790.5040) and infrared (1697 cm -1 for carbamate).
With our chiral carbamate in hand, we attempted a selective bromination reaction using the protocol without palladium, which returned a good yield (>80 %) for the inseparable brominated products 7a and 7b. We had been hoping to use 1 H NMR spectroscopy to quantify the diastereoselectivity of the reaction, but to our disappointment, there were really no promising signals to work with. Different solvents (CDCl 3 , DMSO-d 6 and C 6 H 6 ) and different temperatures (up to the maximum allowed) all failed to help us determine the diastereoselectivity. The only signal that appeared marginally useful was the methine signal on the chiral centre of the menthyl group. In the starting material this appeared at δ 4.60 ppm as a triplet of doublets, whilst in the product it appeared as a multiplet centred around δ 4.69 ppm (see Fig. 3). On close inspection, it could be seen that this multiplet was an overlay of two triplets of doublets in essentially a 1:1 ratio, suggesting a poor diastereoselectivity in the reaction.
In order to see if we could improve on this diastereoselectivity, we tried the reaction at lower temperatures. Since DCE is relatively limited in this sense, we changed the solvent to dichloromethane (DCM) and initiated a temperature study (Table 3). Unfortunately, in all cases (down to -35°C) we saw no discernible improvement in the diastereoselectivity (as judged by the aforementioned signals in the 1 H NMR spectra). However, what was unexpected was just how well the reaction occurred even at lower temperatures, albeit with slightly longer reaction times.
Our inability to determine the ratio of diastereomers in this reaction was frustrating. Even normal and reversed-phased HPLC experiments, including the use of a chiral column, failed to separate the two diastereomers. As a last resort, we decided to remove the chiral menthyl group and examine the optical rotation of the resultant mixture of enantiomers to see if any optical activity was displayed. Although this would not be a means of determining enantioselectivity, the optical rotation would at least point to whether any chiral induction had taken place. Removal of the menthyl group was readily achieved using Coudert's method of tetra-n-butylammonium fluoride (Bu 4 NF, TBAF) in THF (Scheme 3). 24 The reaction, as expected, 24 was sluggish and required heating under reflux for 36 h. Nevertheless, after work-up and purification, the aminocalix [4]arene product was obtained in yields >80 %. The 1 H NMR spectrum showed the complete removal of the menthyl group, greatly simplifying the spectrum. The loss of the carbamate was also detected by IR spectroscopy and the HRMS returned the expected molecular ion (and isotopic distribution pattern) for the product. Optical rotation experiments were then run on material generated from bromination at -35°C (Table 3, entries 5 and 6), returning values of [α] D = +6°and +3°for products derived from entries 5 and 6, respectively. Whilst these values cannot be used for any form of quantification, they do indicate a level of enantiomeric excess, which in turn, points to at least some degree of diastereoselectivity induced by the chiral menthyl carbamate.

Conclusion
In conclusion, we have shown that a report in the literature claiming a C-H activation route, mediated by a catalytic palladium in which a carbamate directs the formation of an aryl halide bond, is somewhat overstated and incorrect on a number of points claimed. Nevertheless, using a chiral calix [4]arene carbamate, bromination successfully delivered a product that suggested a modest level of inherent chirality that could not be quantified. Further work can potentially look at other chiral groups and also at extending the number of directing groups on the upper-rim of the calix [4]arene to two or even four, in order to access more interesting meta-functionalized calix [4]arenes.

Supplementary Material
Copies of NMR, IR and HRMS spectra for all new compounds synthesized and also references for halogenation reactions prior to 2016 are provided in the supplementary material appended to the end of this article.

Figure 3
Comparison of the chiral C-H methine in the brominated 7a and 7b (a) and starting material 6 (b) menthyl carbamate calix [4]arenes (spectra run in CDCl 3 ). gen. Other reagents that required purification were done so according to standard procedures. The synthesis of methyl (4-methoxyphenyl)carbamate 1 was performed using a literature procedure from p-anisidine, 25 and mono-aminocalix [4] arene 3 was prepared as previously reported by us. 22 For syntheses performed under inert conditions the glassware was oven-dried and then placed under vacuum of <0.5 mm Hg before being periodically flushed with argon until reaching room temperature. All reactions were performed under positive pressure of 2.8 kPa of 5.0 grade argon (Air Products). Low temperature reactions were performed in a Dewar containing ice and acetone (-15°C), solid CO 2 and acetonitrile (-40°C) or solid CO 2 and acetone (-78 °C).
Column chromatography was performed using 230-400 nm silica and thin layer chromatography (TLC) was performed using Macherey-Nagel DC-Fertigfolien ALUGRAM Xtra SIL G/UV254 TLC plates. Visualization of compounds on TLC plates was performed by using a UV lamp or using a cerium ammonium molybdate (CAM) solution followed by heating.
Both 1 H and 13 C NMR spectra were obtained using Varian 300 MHz VNMRS, Varian 400 MHz Unity INOVA and Varian 600 MHz Unity INOVA NMR instruments. Chemical shifts were recorded using the residual solvent peaks (chloroform-d or DMSO-d 6 ) and reported in ppm. Unless otherwise stated, NMR spectra was obtained at room temperature. All mass spectrometry spectra were obtained by Central Analytical Facility (CAF) at Stellenbosch University using a Waters API Q-TOF Ultima mass spectrometer. IR spectra were obtained using a Thermo Nicolet Nexus FTIR instrument using the ATR attachment. Melting points were obtained using a Gallenkamp Melting Point Apparatus.

Methyl (2-bromo-4-methoxyphenyl)carbamate (2)
An oven-dried Schlenk equipped with a magnetic stir bar and flushed with argon was charged with 1 (100 mg, 0.552 mmol), NBS (108.5 mg, 1.1 eq), PTSA (48 mg, 0.5 eq) and DCE (1.1 mL). The contents were then heated to 60°C and left to stir for 2.5 h. After 2.5 h, the reaction contents were cooled to room temperature and then diluted with DCM (20 mL). The solution was then poured into H 2 O (20 mL) after which the product was extracted with DCM (10 mL × 3). The organic layers were subsequently combined and washed with a 10 % HCl solution (20 mL), followed by sat. NaHCO 3 (20 mL) solution and finally brine (20 mL). The solution was then dried over MgSO 4 and the solvent was removed via reduced pressure. Purification was achieved via silica gel flash column chromatography (EtOAc: PET 10:90) to obtain compound 2 as an orange solid in 69 % yield (99 mg).