Low-cost instant CO generation at room temperature using formic acid, mesyl chloride and triethylamine

Abstract

Three low-cost standard lab chemicals, formic acid, mesyl chloride and triethylamine, were used for instant carbon monoxide (CO) generation at room temperature. Subsequently this gas was implemented in palladium-catalysed aminocarbonylation chemistry. Moreover, ¹³C-enriched formic acid was used as one of the most economical CO precursors for ¹³C-carbonyl labelling.

---

Carbon monoxide is beyond doubt one of the most important C1 building blocks for organic synthesis.¹ In addition, CO can also act as an excellent ligand in transition metal chemistry.^2–4 Although carbonylation chemistry is very mild and versatile, safety issues constrain the direct utility of this highly toxic gas. Especially on lab scale in a typical research lab, storage and use of pure, pressurised carbon monoxide gas raise serious safety concerns. In order to improve the applicability of this gas, an alternative carbon monoxide source is highly desired.

Different CO sources have successfully been used: examples are metal carbonyl derivatives,⁵ aldehydes,^6,7 formamides,^8,9N-formyl saccharin^10,11 and aryl formates.^12,13 Unfortunately, these precursors are either very toxic themselves, require high temperatures for carbon monoxide liberation or generate a nucleophilic by-product that can participate in the carbonylation reaction.

The Skrydstrup group developed three efficient carbon monoxide generating systems: methyldiphenylsilacarboxylic acid,¹⁴ 9-methyl-9H-fluorene-9-carbonyl chloride¹⁵ and tetraphenyldimethyldisilane in combination with carbon dioxide.¹⁶ These CO-precursors, however, are non-trivial specialty chemicals with an associated higher cost. It would be advantageous if inexpensive commodities could be used to release carbon monoxide at ambient temperature in a robust manner. Different groups are working towards optimising such precursor systems.

In 2015, rapid hydrolysis of chloroform to carbon monoxide by caesium hydroxide hydrate was described.¹⁷ As chloroform is an inexpensive bulk chemical, it is an interesting CO precursor with the possibility of ¹³C labelling since ¹³C-CHCl₃ is commercially available. On the other hand, the heterogeneous hydroxide (CsOH·H₂O) base is an expensive chemical and must be used in excess to ensure rapid hydrolysis of chloroform. Replacement of this caesium salt by a more cost efficient one leads to a drastic decrease in carbon monoxide production.

Generation of carbon monoxide by dehydration of formic acid¹⁸ in the presence of sulfuric acid is known as the Morgan reaction.¹⁹ The utility of this reaction was improved by adapting this CO-generating system to a continuous microflow system by using a tube-in-tube reactor or by making use of a two-chamber system.²⁰ However, this system still requires corrosive concentrated sulfuric acid. Zeolites are known to be safe alternatives to strong and corrosive liquid acids.²¹ Although replacement of sulfuric acid by zeolites seems promising, only a few specific zeolites can adequately decompose formic acid at an elevated temperature (150 °C), hampering the robustness of the method.²²

Recently, the Wu group discovered that formic acid releases carbon monoxide in the presence of acetic anhydride under basic conditions.^23–25 Moderate to excellent yields were obtained when using this CO precursor in Pd-catalysed alkoxycarbonylation, carbonylative Suzuki and Sonogashira coupling of aryl halides. Unfortunately, the decomposition of the mixed anhydride is slow, while instant carbon monoxide generation is desired as it directly stabilises the catalytic system and shortens the reaction time.²⁶ Furthermore, this method is not suitable for Pd-catalysed aminocarbonylation chemistry as the in situ formed acetic formic anhydride is known to be a formylating agent for amines.²⁷

In this paper we report a state-of-the-art low-cost CO generating system based on the instant decomposition of formic acid by the addition of mesyl chloride (MsCl) and triethylamine (Et₃N) at room temperature (Scheme 1). This system is 30 to 140 times more cost efficient than non formic acid precursors (see the ESI†), since all the required starting products are inexpensive commodity chemicals. The replacement of acetic anhydride by mesyl chloride was necessary to ensure instant CO generation at room temperature (see the ESI†). This replacement also contributes to the general safety aspects of the procedure, since mesyl chloride is a non-flammable chemical compared to acetic anhydride. As ¹³C-formic acid is also commercially available, this procedure provides an obvious source of ¹³CO. Comparing the costs of commercially available ¹³CO precursors, this system is probably one of the least expensive ¹³C-carbonylation methods (see the ESI†).

Scheme 1 Proposed decomposition mechanism of formic acid in the presence of 1 equiv. mesyl chloride and 2 equiv. triethylamine leading to instant CO generation.

To gain more insight into the steps involved in the decomposition of formic acid to carbon monoxide in the presence of mesyl chloride and triethylamine, a ¹³C-NMR study was conducted at room temperature (Scheme 2). First, 0.5 mmol of HCOOH was added to 600 μL of CD₃CN (Scheme 2[A]). Then, 0.5 mmol of MsCl was added to this solution (Scheme 2[B]). At this point no changes seemed to occur in the NMR spectrum of the mixture, indicating that a base is needed to start the reaction. Subsequent dropwise addition of 1.0 mmol of Et₃N resulted in vigorous gas development for several seconds. Afterwards the ¹³C-NMR spectrum was recorded (Scheme 2[C]) and complete decomposition of HCOOH was observed as the corresponding peak disappeared. MsCl also disappeared in the spectrum and a new peak at a lower ppm-value appeared (39.6 ppm), indicating the formation of the methanesulfonate anion and the triethylammonium cation (see the ESI†).

Scheme 2 ¹³C-NMR study of the CO precursor at room temperature. [A] 0.5 mmol of HCOOH in 600 μL of CD₃CN. [B] 0.5 mmol of HCOOH and 0.5 mmol of MsCl in 600 μL of CD₃CN. [C] 0.5 mmol of HCOOH, 0.5 mmol of MsCl and 1.0 mmol of Et₃N in 600 μL of CD₃CN.

These results indicate that once the formic acid is deprotonated, it most probably reacts with mesyl chloride and forms a highly unstable mixed anhydride intermediate. A second deprotonation leads to instant CO formation, as the methanesulfonate is an excellent leaving group (Scheme 1). At this point it is not known whether the elimination of the sulfonate happens concerted with the deprotonation or not.

We also found that other sulfonyl chlorides are equally able to decompose formic acid to carbon monoxide in the presence of triethylamine: tosyl chloride, triflyl chloride, nosyl chloride, etc. However, for reasons of atom economy, mesyl chloride was our reagent of choice.

Palladium-catalysed aminocarbonylation of aryl bromides was chosen as the test case to prove the utility of this CO precursor. As already mentioned, in situ formation of CO by e.g. acetic formic anhydride will lead to formylation of the amine. More generally, decomposition of any precursor will lead to the formation of by-products which could potentially cause side reactions. In our case, the amine moiety (needed to form an amide) could react with mesyl chloride and a mixture of compounds might be expected. Inspired by the elegant two-chamber setup of the Skrydstrup group, ex situ CO generation was implemented to address these difficulties.²⁸

We began our optimisation by using the conditions reported by Buchwald et al. as a benchmark for the Pd-catalysed aminocarbonylation chemistry.²⁹ Initially, chamber A was filled with 5 mol% Pd(OAc)₂/Xantphos, Na₂CO₃ (3 equiv.), 3 mL of dry degassed toluene (0.167 M), bromobenzene (1 equiv.) and n-hexylamine (1.5 equiv.). Chamber B was loaded with 3 equiv. formic acid and mesyl chloride. Finally, Et₃N (6 equiv.) was added by injection through the septum in chamber B at room temperature (Table 1). Performing the reaction at 60 °C resulted in 41% of the desired amide 3a after 18 hours (entry 1). To our delight an increased yield was observed at higher temperatures (entries 2 and 3). In contrast to our results at 100 °C, Buchwald reported an incomplete conversion of the starting material at this temperature. This unusual result was ascribed to decreased catalyst stability at higher temperature.²⁹ However, in our system, CO pressure builds up in the closed system. We assume that the generated pressure stabilises the catalytic system and therefore higher temperatures and lower catalyst loadings can be applied.

Table 1 Optimisation of the reaction conditions^a

Entry	T (°C)	t (h)	CO eq.	Catalyst/ligand	Conc. (M)	Yield^b (%)
a Reaction conditions: chamber A – bromobenzene (0.5 mmol), n-hexylamine (0.75 mmol, 1.5 equiv.), Na2CO3 (1.5 mmol, 3.0 equiv.), Pd(OAc)2 (0.005 mmol), Xantphos (0.005 mmol) and 1 mL of dry degassed toluene; chamber B – formic acid (0.65 mmol, 1.3 equiv.), methanesulfonyl chloride (0.65 mmol, 1.3 equiv.) in 2 mL of dry degassed toluene. Finally, Et3N (1.3 mmol, 2.6 equiv.) was added by injection through the septum in chamber B at room temperature. After 2 minutes the reactor was placed in an oil-bath at 100 °C. b Isolated yield.
1	60	18	3.0	5 mol% Pd(OAc)2/Xantphos	0.167	41
2	80	18	3.0	5 mol% Pd(OAc)2/Xantphos	0.167	80
3	100	18	3.0	5 mol% Pd(OAc)2/Xantphos	0.167	96
4	100	18	2.0	5 mol% Pd(OAc)2/Xantphos	0.167	95
5	100	18	1.3	5 mol% Pd(OAc)2/Xantphos	0.167	93
6	100	18	1.3	2 mol% Pd(OAc)2/Xantphos	0.167	96
7	100	18	1.3	1 mol% Pd(OAc)2/Xantphos	0.167	94
8	100	18	1.3	1 mol% Pd(P(Ph3))4	0.167	19
9	100	2	1.3	1 mol% Pd(OAc)2/Xantphos	0.167	95
10	100	2	1.3	1 mol% Pd(OAc)2/Xantphos	0.250	96
11	100	2	1.3	1 mol% Pd(OAc)2/Xantphos	0.500	98

---

In order to increase the cost efficiency of this system, a reduction of CO and catalyst loading was investigated (entries 4–7). Performing the reaction with only 1 mol% Pd(OAc)₂/Xantphos in the presence of 1.3 equiv. CO (assuming full conversion of the precursor) affords the amide in 94%. Further improvements were made by reducing the reaction time and increasing the concentration to 2 hours (entry 9) and 0.5 M (entry 11), respectively.

The ex situ generated CO was successfully implemented in the system comprising bromobenzene and n-hexylamine yielding 3a in 98%. With the optimised conditions in hand, the scope of the Pd-catalysed aminocarbonylation chemistry was further explored. As is shown in Scheme 3, different benzamides and Weinreb amides were synthesised in good to excellent yields.

Scheme 3 Palladium-catalysed aminocarbonylation by using the formic acid, mesyl chloride and triethylamine system as the CO source. ^a Reaction conditions: chamber A – bromobenzene (0.5 mmol), amine (0.75 mmol, 1.5 equiv.), Na₂CO₃ (1.5 mmol, 3.0 equiv.), Pd(OAc)₂ (0.005 mmol), Xantphos (0.005 mmol) in 1 mL of dry degassed toluene; chamber B – HCOOH (0.65 mmol, 1.3 equiv.) and MsCl (0.65 mmol, 1.3 equiv.) in 2 mL of dry degassed toluene. Finally, Et₃N (1.3 mmol, 2.6 equiv.) was added by injection through the septum in chamber B at room temperature. After 2 minutes the reactor was immersed in an oil-bath at 100 °C.

The reaction conditions tolerate a wide variety of functional groups on both the aryl bromide and the amines, including allyl (3c), nitrile (4a and 4e), nitro (4d), aryl chlorides (4c and 6a) and esters (3h). In addition, also heteroaryl bromides (thiophene 4b and pyridine 7a) gave good to excellent yields.

This work was finalised with the synthesis of three relevant ¹³C-labelled pharmaceuticals: CX-546³⁰ (5a and 5b), Moclobemide³¹ (6a and 6b), and Nikethamide³² (7a and 7b). These compounds were labelled using ¹³C-formic acid. Notably, this enriched precursor is at least 9 times less expensive per mmol of ¹³CO generated than any other reported commercial ¹³CO precursor (see the ESI†).

In conclusion, we report a state-of-the-art low-cost system to instantly generate CO at room temperature, which is composed of formic acid, mesyl chloride and triethylamine. Since these reagents are inexpensive commodity chemicals, which are available in research labs, this system belongs to one of the most economical and readily available CO precursor systems. The ex situ generated CO is successfully applied in Pd-catalysed aminocarbonylation chemistry resulting in high yielding amide formation. Remarkably, this is the first report on ¹³C-carbonylation labelling in a Pd-catalysed reaction by decomposition of ¹³C-HCOOH into ¹³CO.

Acknowledgements

CV and BE thank the FWO (PhD Fellowship of Research Foundation – Flanders) for fellowships received. WDB thanks KU Leuven for financial support via project OT/14/067.

Notes and references

1. I. Cipres, J. Jenck and P. Kalck, J. Mol. Catal., 1990, 58, 387–392.
2. G. Kiss, Chem. Rev., 2001, 101, 3435–3456.
3. A. Brennfuhrer, H. Neumann and M. Beller, Angew. Chem., Int. Ed., 2009, 48, 4114–4133.
4. C. J. Mallia and I. R. Baxendale, Org. Process Res. Dev., 2015 DOI:10.1021/acs.oprd.5b00222.
5. A. Wieckowska, R. Fransson, L. R. Odell and M. Larhed, J. Org. Chem., 2011, 76, 978–981.
6. T. Morimoto, K. Fuji, K. Tsutsumi and K. Kakiuchi, J. Am. Chem. Soc., 2002, 124, 3806–3807.
7. T. Morimoto, K. Yamasaki, A. Hirano, K. Tsutsumi, N. Kagawa, K. Kakiuchi, Y. Harada, Y. Fukumoto, N. Chatani and T. Nishioka, Org. Lett., 2009, 11, 1777–1780.
8. K. Hosoi, K. Nozaki and T. Hiyama, Org. Lett., 2002, 4, 2849–2851.
9. T. Ueda, H. Konishi and K. Manabe, Org. Lett., 2012, 14, 5370–5373.
10. T. Ueda, H. Konishi and K. Manabe, Angew. Chem., Int. Ed., 2013, 52, 8611–8615.
11. T. Ueda, H. Konishi and K. Manabe, Org. Lett., 2013, 15, 5370–5373.
12. T. Fujihara, T. Hosoki, Y. Katafuchi, T. Iwai, J. Terao and Y. Tsuji, Chem. Commun., 2012, 48, 8012–8014.
13. N. Alonso, J. D. Munoz, B. Egle, J. L. Vrijdag, W. M. De Borggraeve, A. de la Hoz, A. Diaz-Ortiz and J. Alcazar, J. Flow Chem., 2014, 4, 105–109.
14. S. D. Friis, R. H. Taaning, A. T. Lindhardt and T. Skrydstrup, J. Am. Chem. Soc., 2011, 133, 18114–18117.
15. P. Hermange, A. T. Lindhardt, R. H. Taaning, K. Bjerglund, D. Lupp and T. Skrydstrup, J. Am. Chem. Soc., 2011, 133, 6061–6071.
16. C. Lescot, D. U. Nielsen, I. S. Makarov, A. T. Lindhardt, K. Daasbjerg and T. Skrydstrup, J. Am. Chem. Soc., 2014, 136, 6142–6147.
17. S. N. Gockel and K. L. Hull, Org. Lett., 2015, 17, 3236–3239.
18. R. E. DeRight, J. Am. Chem. Soc., 1933, 55, 4761–4764.
19. J. S. Morgan, J. Chem. Soc., Trans., 1916, 109, 274–283.
20. C. Brancour, T. Fukuyama, Y. Mukai, T. Skrydstrup and I. Ryu, Org. Lett., 2013, 15, 2794–2797.
21. M. J. Climent, A. Corma and S. Iborra, Chem. Rev., 2011, 111, 1072–1133.
22. P. Losch, A. S. Felten and P. Pale, Adv. Synth. Catal., 2015, 357, 2931–2938.
23. X. X. Qi, L. B. Jiang, C. L. Li, R. Li and X. F. Wu, Chem. – Asian J., 2015, 10, 1870–1873.
24. L.-B. J. Xinxin Qi, Hao-Peng Li and Xiao-Feng Wu, Chem. – Eur. J., 2015, 21, 17650–17656.
25. X. Qi, C.-L. Li, L.-B. Jiang, W.-Q. Zhang and X.-F. Wu, Catal. Sci. Technol., 2016 10.1039/C5CY01957E.
26. F. Aubke and C. Wang, Coord. Chem. Rev., 1994, 137, 483–524.
27. W. Stevens and A. Vanes, Recl. Trav. Chim. Pays-Bas, 1964, 83, 863–872.
28. R. H. Taaning, P. Hermange, A. T. Lindhardt, S. D. Friis and T. Skrydstrup, Denmark Pat., PA201070546, 2010.
29. J. R. Martinelli, D. A. Watson, D. M. M. Freckmann, T. E. Barder and S. L. Buchwald, J. Organomet. Chem., 2008, 73, 7102–7107.
30. T. Lipina, K. Weiss and J. Roder, Neuropsychopharmacology, 2007, 32, 745–756.
31. G. Laux, Psychiatr Prax., 1989, 16, 37–40.
32. O. Feinsilver, Curr. Ther. Res. Clin. Exp., 1962, 4, 165–177.

---

Footnotes

† Electronic supplementary information (ESI) available: ¹³C-NMR study of the CO generating system, influence of MsCl and Ac₂O on the decomposition rate of HCOOH, price comparison of recently published CO and ¹³CO generating systems, detailed experimental procedures and copies of ¹H-NMR and ¹³C-NMR data. See DOI: 10.1039/c6re00006a

‡ Authors contributed equally.

---