Abstract
Aldehyde decarbonylation is a vital chemical transformation in the synthesis of natural products. Nature accomplishes this process through a family of decarbonylase enzymes, while in the laboratory, harsh transition metals and elevated temperatures are required. Herein, we report a mild aldehyde decarbonylation reaction that exhibits exclusive selectivity for ortho-aldehydes during a tandem nitrile boronic acid cross-coupling reaction. A wide substrate scope is displayed that includes regioselective removal of the ortho-aldehyde from phenyl boronic acids in the presence of meta- or para-aldehydes. A mechanistic investigation of the observed regioselectivity for ortho-aldehydes by density functional theory (DFT) calculations shows that the CO ligand extrusion is energetically more favorable for the ortho position as compared to the para position.
Keywords: decarbonylation, aldehydes, boronic acids, nitriles, palladium
Graphical Abstract

Aldehyde decarbonylation involving the loss of CO was first developed by Eschinazi and co-workers in the 1950s,1 and has been a crucial transformation in the synthesis of natural products. Aldehydes are often introduced during total synthesis as handles to promote improved reactivity for numerous organic reactions including Diels–Alder,2 oxa-Michael additions,3 C–H activations,4 and many more.5 However, the removal of aldehydes is challenging and often requires harsh conditions and metals including Rh,6 Ni,7 Ru,8 Ir,9 and Pd.10 More recently, nanoparticles,11 photocatalysts,12 and microwave conditions13 have been employed to overcome the high cost and harsh conditions associated with these catalysts. Additionally, many reports utilize an aldehyde scavenger to minimize catalyst poisoning.14 Yet, despite its high cost and use in stoichiometric amounts, Wilkinson’s Rh catalyst remains the reagent of choice for this transformation.15
Regardless of these numerous methods for aldehyde decarbonylation, none are capable of differentiating various aldehydes and indiscriminately remove CO from all formyl groups, often requiring chemists to mask desired aldehydes in different oxidation states, lengthening synthetic routes.16
While investigating the substrate scope of a cross-coupling reaction between nitrile 1a and an aryl boronic acid 2a using Pd(OAc)2,17 we serendipitously noticed the disappearance of the aldehyde group (Table 1, entry 1). Encouraged by the one-pot nature of this aldehyde decarbonylation, we sought to explore this reaction further. Increasing the temperature from 60 °C to 90 °C decreased the reaction time from 48 hours to 12 hours without significant change in the yield of the reaction (Table 1, entries 2 and 3). Increasing the catalyst loading to 20 mol% Pd(OAc)2 with 40 mol% bipyridine ligand did not produce any significant difference in the yield or reaction time (Table 1, entries 4 and 5). All reactions were purged under N2 to mitigate unproductive oxidative homocoupling of the phenyl boronic acid, which hinders reaction progress.18
Table 1.
Aldehyde Decarbonylation of Formyl Phenyl Boronic Acid (2a) via a Nitrile Boronic Acid Cross-Coupling Reactiona
| |||||
|---|---|---|---|---|---|
| Entry | Pd(OAc)2 (mol%) | bpy (mol%) | Temp (°C) | Time (h) | Yielda |
| 1 | 10 | 20 | 60 | 48 | 72% |
| 2 | 10 | 20 | 80 | 24 | 81% |
| 3 | 10 | 20 | 90 | 12 | 74% |
| 4 | 20 | 40 | 80 | 24 | 79% |
| 5 | 20 | 40 | 90 | 12 | 72% |
All reactions were conducted with 50 mg of 1a, 4 equiv. of 2a, and 10 equiv. of TFA in 5:1 THF:H2O.
Yield refers to chromatographically pure compounds.
Aldehyde decarbonylation using Pd(OAc)2 has been reported previously by Maiti and collaborators,10f but they required high temperatures up to 140 °C, while we observed the decarbonylation of aldehydes at lower temperatures of 60 °C. Interestingly, the cross-coupling reaction with meta-formyl phenyl boronic acid (2b) did not lead to any decarbonylation under the reaction conditions, only generating ketone 3b in high yield (76%) (Scheme 1). Literature regarding regioselective decarbonylation is absent, leading us to further investigate this unique reactivity. Similar to the meta-substituted aldehyde, no aldehyde decarbonylation occurred with para-formyl phenyl boronic acid (2c) and only cross-coupled product 3c (71%) was observed under the reaction conditions. This observation motivated us to further explore the substrate scope of this regioselective aldehyde decarbonylation reaction. To determine the scope and role of substituents in ortho-selective aldehyde decarbonylation, we explored substrates with electron-donating (2d–f) and electron-withdrawing (2g–i) groups at varying positions with respect to the aldehyde handle. All the reactions successfully underwent aldehyde decarbonylation with nitrile 1a under the developed reaction conditions and yielded ketone products 3d–i (Scheme 1).
Scheme 1.

Scope of boronic acids in the regioselective ortho-aldehyde decarbonylation reaction. Yields refer to chromatographically pure compounds. a 12 hours. No decarbonylation was observed with substrates containing an aldehyde group at the meta (2b) and para (2c) positions of the phenyl boronic acid. b 48 hours.
ortho-formyl phenyl boronic acids containing electron-withdrawing substituents, particularly trifluoromethyl, gave modestly lower yields after 48 hours (e.g., 3i, 45%) (Scheme 1). We mainly attribute the lower conversion to the slower rate of the initial oxidative addition with the nitrile group.17 However, the reaction does require an aromatic boronic acid, as vinylic boronic acids gave no reaction.
Next, we evaluated the substrate scope of this reaction with various nitriles (1b–g) by reaction with ortho-formyl phenyl boronic acid (2a) (Scheme 2). Phenyl nitriles 1b–d and naphthyl nitrile 1e yielded the decarbonylated ketone products 3j–m in good yields (64–84%). An electron-withdrawing functionality on the nitrile 1f was well tolerated and generated the decarbonylated product 3n in high yield (70%) under the reaction conditions. Most notably, we were able to selectively decarbonylate the ortho-aldehyde of the phenyl boronic acid 2a in the presence of meta- and para-aldehydes on the phenyl nitrile 1b,c to generate the corresponding ketone products 3j,k in high yields. To the best of our knowledge, these are the first examples of regioselective decarbonylation. Additionally, a reactive carbonyl moiety was retained, as seen with the ester group in 3o. Interestingly, nitrile substrates with β-carbonyl functionality were not tolerated.
Scheme 2.

Scope of nitriles in the regioselective ortho-aldehyde decarbonylation reaction. Yields refer to chromatographically pure compounds.
In order to investigate the role of the boronic acid in aldehyde decarbonylation, we attempted the reaction with 2-acetylbenzaldehyde as a model substrate and subjected it to the optimized conditions (Scheme 3). No decarbonylation was observed and starting material 2-acetylbenzaldehyde was fully recovered under the conditions. We then postulated that the boronic acid may be supporting the coordination of palladium to facilitate the decarbonylation. However, the addition of 2-formylphenylboronic acid or boric acid externally in the presence of 2-acetylbenzaldehyde did not lead to any decarbonylation (Scheme 3). These experiments suggest that the ortho decarbonylation stems from the cross-coupling reaction of the nitrile with the boronic acid in a concerted manner.
Scheme 3.

Control reactions to investigate the mechanistic pathway. Reactions were conducted on 50 mg scale of 2-acetylbenzaldehyde at 80 °C for 24 hours. No conversion was observed for all conditions.
Based on these experimental results and literature precedent for the nitrile cross-coupling reaction19 and aldehyde decarbonylation,20 we proposed a plausible reaction mechanism (Scheme 4A). In the first step, Pd(OAc)2 coordinates with bipyridine and trifluoroacetate to generate the active Pd(II) catalyst A, which further undergoes transmetalation with aryl boronic acid 2, resulting in the Pd-aryl species B (steps i and ii, Scheme 4A). The coordination of nitrile 1 with Pd-aryl species B generates intermediate C, which is followed by intramolecular carbopalladation to form the corresponding imine Pd(II) complex D before undergoing reductive elimination, yielding imine complex E (steps iii–v, Scheme 4A). The oxidative addition of Pd into the ortho-formyl C–H bond generates intermediate F, after which migratory deinsertion gives complex G (steps vi and vii, Scheme 4A). This is followed by reductive elimination, resulting in the generation of decarbonylated imine product H and a Pd catalyst CO complex (step viii, Scheme 4A). Imine H further undergoes hydrolysis to generate the decarbonylated ketone product 3, while Pd dissociates from CO to regenerate the active catalyst A with trifluoroacetate ligands (step ix, Scheme 4A).
Scheme 4.

Reaction Pathway. (A) Proposed mechanistic pathway. (B) Energy landscape. Values are in kcal/mol relative to the Ortho E intermediate. Computations conducted at B3LYP/LanL2DZ theory and basis set.
Given the high regioselectivity for ortho-aldehydes, we decided to investigate a plausible explanation using density functional theory (DFT) calculations (B3LYP/LanL2DZ) on ortho- and para-imine intermediates E–H, where R = CH3. Scheme 4B displays the energies in kcal/mol of each intermediate relative to the Ortho H intermediate. While the C–H oxidative addition on Ortho E and Para E intermediates (step vi, forming intermediates Ortho F and Para F) is unfavorable for both substrates, the CO ligand extrusion is a favorable process (step vii) for the Ortho F system, generating Ortho G, which has ~20 kcal/mol lower energy than Ortho F. Conversely, the transformation from Para F to generate Para G is unfavorable by ~17 kcal/mol. Overall, the CO ligand extrusion process is the rate-determining step,20 and is ~37 kcal/mol more favorable for the formation of Ortho G as compared to Para G (step vii). We postulated that Ortho G is likely stabilized by the electron donation from the imine in a 5-membered transition state, giving palladium 18 electrons, the preferred electronic configuration for palladium.21 These computations are in agreement with the excellent regioselectivity of this method, compared to the traditional global decarbonylation approaches.
In summary, we report a regioselective aldehyde decarbonylation, the first to our knowledge, via a nitrile boronic acid cross-coupling reaction under mild conditions.22 On the basis of experimental and computational analysis, we concluded that the exclusive ortho selectivity is due to the lower activation energy for the decarbonylation of ortho-aldehydes over meta- or para-aldehydes. The substrate scope was explored with a variety of boronic acids and diverse nitriles, demonstrating the broad compatibility of this reaction. This method exhibits the ability to transform and remove functionality in a one-pot manner under mild conditions with low catalyst loading, while demonstrating the potential for other directing groups to promote selective decarbonylation.
Supplementary Material
Acknowledgment
We thank the Emory Mass Spectrometry Center for HRMS analysis.
Funding Information
This research was supported by National Institutes of Health (NIH) grants (1R35GM133719-01 and 1R01HG012941-01) to M.R. M.R. was also supported by a Research Scholar Grant (RSG-22-025-01-CDP) from the American Cancer Society.
Footnotes
Conflict of Interest
The authors declare no conflict of interest.
Supporting Information
Supporting information for this article is available online at https://doi.org/10.1055/s-0042-1751562.
References and Notes
- (1).(a) Eschinazi HE Bull. Soc. Chem. Fr 1952, 967. [Google Scholar]; (b) Eschinazi HE; Pines H J. Org. Chem 1959, 24, 1369. [Google Scholar]
- (2).(a) Jones SB; Simmons B; Mastracchio A; MacMillan DWC Nature 2011, 475, 183. [DOI] [PMC free article] [PubMed] [Google Scholar]; (b) Berrué F; McCulloch MWB; Kerr RG Bioorg. Med. Chem 2011, 19, 6702. [DOI] [PubMed] [Google Scholar]
- (3).Hong B; Kotame P; Tsai C-W; Liao J-H Org. Lett 2010, 12, 776. [DOI] [PubMed] [Google Scholar]
- (4).Goswami N; Bhattacharya T; Maiti D Nat. Rev. Chem 2021, 5, 646. [DOI] [PubMed] [Google Scholar]
- (5).Lu H; Yu T-Y; Xu P-F; Wei H Chem. Rev 2021, 121, 365. [DOI] [PubMed] [Google Scholar]
- (6).(a) Tsuji J; Ohno K Tetrahderon Lett. 1967, 8, 2173. [Google Scholar]; (b) Doughty DH; Pignolet LH J. Am. Chem. Soc 1978, 100, 7083. [Google Scholar]; (c) Kries M; Palmelund A; Bunch L; Madsen R Adv. Synth. Catal 2006, 348, 2148. [Google Scholar]
- (7).(a) Bock H; Breuer O Angew. Chem., Int. Ed. Engl 1987, 26, 461. [Google Scholar]; (b) Ding K; Xu S; Alotaibi R; Paudel K; Reimheiner EW; Weatherly J J. Org. Chem 2017, 82, 4924. [DOI] [PubMed] [Google Scholar]
- (8).(a) Domazetis G; Tarpey B; Dolphin D; James BR J. Chem. Soc., Chem. Commun 1980, 939. [Google Scholar]; (b) Bora BR; Prakash R; Sultana S; Gogoi S Org. Biomol. Chem 2021, 19, 2725. [DOI] [PubMed] [Google Scholar]
- (9).(a) Iwai T; Fujihara T; Tsuji Y Chem. Commun 2008, 46, 6215. [DOI] [PubMed] [Google Scholar]; (b) Olsen EPK; Madsen R Chem. Eur. J 2012, 18, 16023. [DOI] [PubMed] [Google Scholar]
- (10).(a) Hawthorne JO; Wilt MH J. Org. Chem 1960, 25, 2215. [Google Scholar]; (b) Tsuji K; Ohno T; Kajimoto T Tetrahedron Lett. 1965, 6, 4565. [Google Scholar]; (c) Tsuji K; Ohno T J. Am. Chem. Soc 1968, 90, 94. [Google Scholar]; (d) Wilt JW; Pawlikowski WW Jr. J. Org. Chem 1975, 40, 3641. [Google Scholar]; (e) Matsubara S; Yokota Y; Oshima YK Org. Lett 2004, 6, 2071. [DOI] [PubMed] [Google Scholar]; (f) Modak A; Deb A; Patra T; Rana S; Maity S; Maiti D Chem. Commun 2012, 48, 4253. [DOI] [PubMed] [Google Scholar]; (g) Modak A; Naveen T; Maiti D Chem. Commun 2013, 49, 252. [DOI] [PubMed] [Google Scholar]
- (11).Kundu PK; Dhiman M; Modak A; Chowdhury A; Polshettiwar V; Maiti D ChemPlusChem 2016, 81, 1142. [DOI] [PubMed] [Google Scholar]
- (12).Kolb D; Morgenstern M; König B Chem. Commun 2023, 59, 8592. [DOI] [PubMed] [Google Scholar]
- (13).Akanksha; Maiti D Green Chem. 2012, 14, 2314. [Google Scholar]
- (14).O’Conner JM; Ma J J. Org. Chem 1992, 57, 5075. [Google Scholar]
- (15).Gutmann B; Elsner P; Glasnov T; Roberge DM; Kappe CO Angew. Chem. Int. Ed 2014, 53, 11557. [DOI] [PubMed] [Google Scholar]
- (16).Selaković Ž; Nikolić AM; Ajdačić V; Opsenica IM Eur. J. Org. Chem 2022, e2102101265. [DOI] [PubMed] [Google Scholar]
- (17).Wang X; Wang X; Liu M; Ding J; Chen J; Wu H Synthesis 2013, 45, 2241. [Google Scholar]
- (18).(a) Xu Z; Mao J; Zhang Y Catal. Commun 2008, 9, 97. [Google Scholar]; (b) Darzi ER; White BM; Loventhal LK; Zakharov LN; Jasti R J. Am. Chem. Soc 2017, 139, 3106. [DOI] [PubMed] [Google Scholar]
- (19).Wang X; Liu M; Xu L; Wang Q; Chen J; Ding J; Wu H J. Org. Chem 2013, 78, 5273. [DOI] [PubMed] [Google Scholar]
- (20).(a) Fristrup P; Kries M; Plamelund A; Norrby P-O; Madsen R J. Am. Chem. Soc 2008, 130, 5206. [DOI] [PubMed] [Google Scholar]; (b) Modak A; Rana S; Phukan A; Maiti D Eur. J. Org. Chem 2017, 4168. [Google Scholar]; (c) Xie H; Qi T; Lyu Y-J; Zhang J-F; Si Z-B; Liu L-J; Zhu L-F; Yang H-Q; Hu C-W Phys. Chem. Chem. Phys 2021, 21, 3795. [DOI] [PubMed] [Google Scholar]
- (21).Sheong FK; Chen W-J; Lin Z J. Organomet. Chem 2015, 792, 93. [Google Scholar]
- (22).Aldehyde Decarbonylation; General Procedure To a mixture of THF:H2O (5:1) in a 35 mL high-pressure tube, TFA (10 equiv.) was added followed by the addition of nitrile 1 (1 equiv.), formyl phenyl boronic acid 2 (4 equiv.), and 2,2′-bipyridyl ligand (40 mol%). The mixture was left to stir until all compounds had dissolved, after which N2 was bubbled for 5 minutes. Finally, Pd(OAc)2 (20 mol%) was transferred to the reaction vessel (order of addition reduces oxidative homocoupling of boronic acid). The high-pressure tube was again flushed with N2 and the contents left to stir for 24 hours at 80 °C. The reaction progress was monitored by TLC. After completion of the reaction, the mixture was diluted with EtOAc (10 mL), extracted with brine (3 × 15 mL), dried over Na2SO4, and evaporated. The crude reaction mixture was then purified by silica gel column chromatography (EtOAc/hexane gradient), and the resulting products were characterized by NMR.
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
