Upscaling of gaseous alkanes into large-volume commodity chemicals via photocatalytic acylation

Summary

From a circular economy point of view, the valorization of gaseous alkanes into large-volume commodity chemicals is of the utmost importance. Such protocols would help reduce the current reliance on a petroleum-based economy and contribute to the challenge of reducing greenhouse gas emissions. In this regard, we hereby report a methodology based on dual photoredox/nickel catalysis that enables the direct coupling of gaseous alkanes with a range of acid chlorides. The protocol is operationally simple, proceeds under mild reaction conditions, and features high chemo- and regioselectivity and good functional group tolerance. Of note, this method serves as an efficient tool for the upscaling of feedstock gaseous alkanes into industrially relevant ketones, such as propiophenone, acetophenone, or isobutyrophenone derivatives.

Graphical abstract

Highlights

• •First acylative cross-coupling of gaseous alkanes
• •Conversion of main components of natural gas into commercially relevant ketones
• •Broad scope along with high chemo- and regioselectivity

Nair et al. report an acylative cross-coupling for the upscaling of gaseous alkanes into commercially relevant ketone derivatives. The method is based on a nickel/photoredox dual-catalysis strategy and facilitates the acylation of compounds such as methane, ethane, and propane in good yields with high chemo- and regioselectivity.

Introduction

Gaseous alkanes (natural gas) are one of the most abundant carbon-based feedstocks on this planet. They have been typically burned off as fuel or used in energy-intensive processes such as steam cracking or reforming.^1,2 However, these processes lack sustainability and heavily contribute to the release of greenhouse gases. Consequently, many international agencies have set a course to completely phase out the burning of natural gas as fuel by 2040.³ As a result, the direct conversion of gaseous alkanes into large-volume commodity chemicals is highly warranted (Scheme 1A). Indeed, if achieved, such methods could lay the foundation for addressing the dual challenge of reducing the current reliance on the petroleum-based economy and contributing to efforts to lower greenhouse gas emissions.^4,5 Although rewarding from a sustainability and efficiency point of view, the functionalization of gaseous alkanes remains challenging due to their extreme lack of reactivity. The C–H bonds in gaseous alkanes have high bond dissociation energy (99–105 kcal mol⁻¹) and ionization potential (∼12.5 eV), along with low acidity (pKa ∼50–51). Also, their gaseous nature and low solubility in common organic solvents add further complications. Specific functionalization of gaseous alkanes has been achieved through oxygenations,⁶ borylations,^7,8 halogenations,⁹ or carbene insertions.^10,11 Furthermore, photocatalytic hydrogen atom transfer (HAT) has emerged in recent years as a viable method for C–H bond activation of gaseous alkanes through the generation of carbon-centered radicals.^12,13,14,15 This technology has been used for the alkylation of electron-deficient alkenes (Giese-type reactions), albeit the resulting products feature limited synthetic applicability.^16,17,18,19 Noël and co-workers in 2023 reported a three-component carbonylative Giese reaction using gaseous alkanes and CO.²⁰ Though highly elegant from a fundamental and operational point of view, the use of highly activated Michael acceptors results in products having limited commercial utility. More recently, our group and that of Noël independently reported protocols based on dual nickel/photoredox catalysis for the arylation of gaseous alkanes with (hetero)aryl bromides in batch and flow, respectively.^21,22 Despite these advances, protocols that allow the upscaling of gaseous alkanes into large-volume commodity chemicals remain underdeveloped.

Scheme 1.

Overview of the acylation of gaseous alkanes

(A) Opportunities in upscaling gaseous alkanes to ketones.

(B) Traditional routes to access propiophenones.

(C) Our hypothesis: direct acylation of ethane to access propiophenones.

(D) This work: photocatalytic direct acylation of gaseous alkanes.

Very few functional groups can rival the practicality offered by ketones in organic synthesis (Scheme 1A).^23,24 For instance, propiophenone (phenyl ethyl ketone) and its derivatives are commodity chemicals having a multibillion-dollar global market, finding utility as industrial solvents, perfumes, flavoring agents, and insect pheromones.²⁵ Moreover, their rich synthetic potential has made them invaluable precursors for the synthesis of several commercial drug molecules, such as phenmetrazine, eprazinone, etoxadrol, tolperisone, and others.^26,27 Industrially, propiophenones are synthesized under harsh reaction conditions (>400°C) via either Friedel-Crafts acylation of arenes or decarboxylative acylation of carboxylic acids (Scheme 1B).^28,29 Alternatively, these ketones are also synthesized at the lab scale using Weinreb amides and Grignard reagents³⁰ or via metal-catalyzed cross-coupling of benzoyl chloride with organometallic reagents.^31,32,33 However, the poor step and atom economies and the use of air-sensitive reagents compromise the sustainability of these methods.

While the direct acylation of liquid alkanes has been reported,^34,35,36 the acylation of more challenging gaseous C1–C4 alkanes has remained elusive to date. We envisioned that the direct acylation of gaseous alkanes with acid chlorides would provide facile access to commercially relevant ketones (Scheme 1C). Such a protocol would not only be highly atom efficient but could offer a promising opportunity for the valorization of gaseous alkanes. In this regard, we hereby report the successful realization of this idea based on a dual nickel/photoredox catalytic strategy. (Scheme 1D). The protocol allows the direct coupling of several gaseous alkanes with a range of aromatic and aliphatic acid chlorides and provides a gateway for the efficient upscaling of gaseous alkanes into industrially relevant ketone derivatives such as propiophenones, acetophenones, isobutyrophenones, or 2-methyl butyrophenones. Furthermore, a curious wavelength-dependent dichotomy was discovered for the acylation of n-butane with benzoyl chloride, wherein the expected 2-methyl butyrophenone is selectively obtained at 390 nm, whereas at 370 nm, propiophenone is obtained as a single product.

Results and discussion

Reaction optimization

We commenced our explorations by studying the acylation of propane with 4-methoxy benzoyl chloride S1, using a dual-catalytic system comprising TBADT as photocatalyst and a nickel co-catalyst (Table 1). Extensive screening of the reaction parameters (supplemental methods: reaction optimization; Tables S1–S6) led to the optimized conditions, which involve a 6 bar pressure of propane, 2 mol % of TBADT, 4 mol % of NiCl₂dtbpy, and K₃PO₄ (2.2 equiv), using MeCN (0.05 M) as a solvent under 44 W 370 nm Kessil lamp irradiation. Under these conditions, product 1 was isolated in an 82% yield as an 11:1 mixture of branched:linear (b:l) regioisomers (entry 1). The use of a lower pressure of propane (3.5 bar) was found to be deleterious, leading to 1 in only a 54% yield (entry 2). The use of NiCl₂dtby was found to be key, as other nickel complexes, such as NiCl₂bpy, NiCl₂dmebpy, and NiCl₂domebpy, provided inferior results (entry 3). The catalyst loading of either TBADT or NiCl₂dtby could be reduced, albeit product 1 was obtained in a slightly diminished yield (entries 4–5). Concentration had a more profound effect on the efficiency of the reaction, as low yields were observed when either more diluted (entry 6) or more concentrated (entry 7) conditions were used. Other bases (entry 8) or light wavelengths (entry 9) proved to be less efficient than the optimal K₃PO₄ and 370 nm. Finally, control experiments revealed that the reaction does not take place in the absence of TBADT, NiCl₂dtbpy, or light irradiation (entry 10).

Table 1.

Optimization table

Entry	Variations from the standard conditions	Yield (%) (b/l ratio)b
1a	none	82 (11:1)
2	3.5 bar of propane	54 (10:1)
3	NiCl2bpy, NiCl2dmebpy, and NiCl2domebpy as [Ni] source	41–60 (7:1-9:1)
4	TBADT (1 mol %)	73 (11:1)
5	NiCl2dtbpy (2 mol %)	72 (11:1)
6	MeCN (0.025 M)	22 (8:1)
7	MeCN (0.1 M)	51 (8:1)
8	2,6-lutidine or collidine instead of K3PO4	50–53 (7:1)
9	44 W 390 nm Kessil lamp	54 (10:1)
10	no TBADT or NiCl2DTBPY or no light	traces

dtbpy, 4,4-di-tert-butyl-2,2-dipyridyl; dbpy, 2,2-dipyridyl; dmebpy, 4,4-dimethyl-2,2-dipyridyl; domebpy, 4,4-dimethoxy-2,2-dipyridyl.

^aStandard reaction conditions: 4-methoxy benzoyl chloride S1 (0.2 mmol), propane (6 bar), TBADT (2 mol %), NiCl₂dtbpy (4 mol %), K₃PO₄ (0.44 mmol), MeCN (4 mL), and 1,200 rpm stirring and irradiation by a 370 nm 44 W Kessil LED lamp for 16 h at room temperature (RT). Full conversion of acid chloride was observed in all cases.

^bYield of isolated product. Regioisomeric ratios (branched vs. linear) are shown in brackets; the structure of the major isomer is depicted.

Substrate scope

After having optimized the acylation of propane, we set out to explore the scope of the reaction (Scheme 2). Initially, we studied the reactivity of various aromatic acid chlorides at 6 bar pressure of propane. Benzoyl chlorides having electron-rich substituents at the para-position underwent facile coupling with propane to afford the corresponding ketones 2–5 in good yields with high branched selectivity (up to 18:1 b:l). Remarkably, electron-neutral and electron-withdrawing groups at the para-position led to the exclusive formation of isobutyrophenones 6–9 (>20:1 b:l), albeit in slightly lower yields. Benzoyl chlorides bearing ortho- (10 and 11) and meta- (12–14) substituents also proved compatible for this transformation. Interestingly, the presence of methyl groups at the meta-position led to the exclusive isolation of the corresponding branched isobutyrophenones 13 and 14. Heteroaromatic acid chlorides could also be used, as illustrated by the synthesis of 15. Notably, aliphatic acid chlorides (linear, branched, and cyclic) also proved to be suitable substrates for this transformation, furnishing products 16–20 in good yields. Unfortunately, N-heterocycles (e.g., quinoline, indole, and pyrrole)-containing acyl chlorides failed to deliver the corresponding products.

Scheme 2.

Substrate scope for the acylation of gaseous alkanes

Reactions run at 0.2 mmol scale using optimized conditions (Table 1, entry 1), unless noted otherwise. Yields refer to isolated products. The regiosiomeric ratio (branched vs. linear) is shown in brackets; the structure of the major isomer is depicted. ^a3 mol % TBADT, 3 mol % [Ni]. ^bMeCN (0.025 M). ^cCD₃CN (0.0125 M). ^d53 W 390 nm Kessil lamp.

Importantly, this photocatalytic methodology could be extended to the acylation of ethane and methane, the main components of natural gas and the most inert alkanes in terms of bond dissociation energy (101 and 105 kcal/mol, respectively). A slight modification in the catalyst loading, along with higher pressure (32 bar) and dilution (0.0125 M), was necessary to efficiently promote the acylation of ethane. Under these modified conditions, a range of benzoyl chlorides underwent facile alkylation with ethane to afford commercially important propiophenones 21–32 in moderate to excellent yields. Notably, the reaction could be scaled up to the 2 mmol scale, as highlighted by the synthesis of 23 (see the supplemental information for details). Not surprisingly, acylation of methane was found to be more challenging and required a higher pressure of 85 bar. Although products were obtained in a low yield, the industrial and commercial relevance of the corresponding acetophenones, 33 and 34, represents a promising avenue for the valorization of this highly abundant gas. Of note, these reactions led to significant amounts (ca. 65% yield) of the corresponding benzoic acids, which could be easily recycled back to the respective benzoyl chloride starting materials. Benzoic acid is likely formed by the hydrolysis of acyl nickel species formed in situ due to the presence of adventitious water, although the direct hydrolysis of acid chloride cannot be fully discarded.

A curious wavelength-dependent dichotomy was unveiled when the acylation of n-butane was studied. 2-Methyl butanones 35–41 were obtained in moderate yields with high regioselectivities from a range of acid chlorides by using 390 nm light irradiation. The change in wavelength was required since, in most cases, the use of a 370 nm light-emitting diode (LED) source led to the selective formation of the corresponding propiophenone derivatives (Scheme 3A). This initially surprising observation can likely be explained by means of a Norrish type II reaction³⁷ (Scheme 3B). Upon photoexcitation at 370 nm, the 2-methyl butyrophenone generated by the acylation of n-butane would generate diradical intermediate A, which undergoes a 1,5-HAT, leading to intermediate B. The ensuing radical fragmentation of B would deliver ethylene and the corresponding propiophenone after tautomerization of intermediate C. These results not only highlight how the chemoselectivity of the acylation of n-butane can be easily tuned by a simple change in LED light irradiation but also offer an interesting alternative for the formation of a highly relevant commodity chemical, such as propiophenone, from another abundant feedstock under rather mild reaction conditions (1 bar, room temperature).

Scheme 3.

Acylation/Norrish type II sequence

(A) Tandem acylation/Norrish type II reaction of n-butane. Reaction conditions: benzoyl chloride (0.2 mmol), n-butane (1 bar), TBADT (2 mol %), NiCl₂dtbpy (4 mol %), K₃PO₄ (0.44 mmol), MeCN (8 mL), and 1,200 rpm stirring and irradiation by a 370 nm 44 W Kessil LED lamp for 16 h at room temperature (RT). Full conversion of acid chloride was observed in all cases. Yields refer to isolated products. ^aThe product ratios (propiophenone vs. 2-methyl butyrophenone) are shown in brackets.

(B) Proposed mechanism of the Norrish type II reaction.

Mechanistic studies

Several experiments were carried out for the acylation of propane to gain insight into the reaction mechanism. The reaction was totally suppressed in the presence of TEMPO (1.5 equiv). TEMPO adducts were detected by gas chromatography-mass spectrometry (GC-MS) in a 7:1 (b:l) ratio, thus demonstrating the intermediacy of both branched and linear alkyl radicals (Scheme 4A). While product 6 was obtained with perfect regioselectivity (b:l > 20:1) at 370 nm, the same reaction at 390 nm delivered an 11:1 b:l mixture of regioisomers in a slightly lower yield (Scheme 4B). This could be attributed to degradation at 370 nm of the minor linear product via a Norrish type II reaction analogous to the one discussed in Scheme 3. The low selectivities observed in some examples (15 and 17–20) could likely be due to the absence of the Norrish type II reaction in those cases. The reactivity and catalytic efficiency of preformed acyl-Ni(II) complex 42 were investigated. No product formation was observed when a stoichiometric amount of 42 was used instead of the acid chloride (Scheme 4C). The formation of alkyl radicals (using TEMPO) was observed in this reaction, indicating that the TBADT-mediated HAT photocatalytic cycle is not hampered by complex 42 (Scheme 4D). These results suggest that the absence of reactivity of 42 might be attributed to its rapid photodegradation in the absence of acid chloride, given the strong light absorption properties of this type of Ni(II) complex.^38,39,40

Scheme 4.

Mechanistic studies

(A) Radical quenching experiment.

(B) Wavelength-dependent regioselectivity.

(C) Reaction using stoichiometric Ni-acyl complex with catalytic TBADT.

(D) Reaction using stoichiometric Ni-acyl complex with TEMPO and TBADT.

(E) Reaction between propane and p-toluyl chloride using catalytic Ni-acyl complex 42.

(F) Hammett studies.

In contrast, Ni(II) complex 42 was found to be catalytically compatible with the acylation of propane, with p-toluyl chloride S3 delivering 3 in a 56% yield, while no formation of product 6 arising from 42 was detected (Scheme 4E). Taken together, these results indicate that the reaction might not involve the direct addition of alkyl radicals onto intermediate 42 to form a Ni(III) complex prior to the C–C-bond-forming event. It is more likely that the key Ni(III) intermediate is generated by the oxidative addition of the acid chloride into the alkyl-Ni(I) complex generated upon the combination of the alkyl radical and the Ni(0) catalyst. Finally, we carried out a Hammett study by comparing the reactivity of electronically variant para-substituted benzoyl chlorides vs. that of benzoyl chloride (Scheme 4F). The observed inverse correlation (log(convX/convH) vs. Hammet constant [σ_p]) indicated that positive charge density might be built up at the carbonyl carbon in the rate-determining step.

On the basis of these studies and literature precedents,^21,22,40,41 the mechanism shown in Scheme 5 could be proposed. The alkyl radical generated by hydrogen abstraction by the photoexcited decatungstate ∗[W₁₀O₃₂]⁴⁻ would combine with the Ni(0) catalyst I (formed by reduction of the Ni(II) precatalyst) to form the alkyl-Ni(I) intermediate II. Subsequent oxidative addition of the acyl chloride into II generates Ni(III) intermediate III, which, upon reductive elimination, furnishes the product and Ni(I) complex IV. Finally, single-electron transfer between IV and the doubly reduced decatungstate [W₁₀O₃₂]⁶⁻2H⁺ (formed by the disproportionation of the singly reduced decatungstate [W₁₀O₃₂]⁵⁻H⁺ generated during the initial hydrogen abstraction) would regenerate the Ni(0) catalyst I, as well as TBADT, closing both catalytic cycles. Besides potential minor product degradation (see above), the high branched selectivity observed for most of the products derived from propane acylation could hint toward a fast oxidative addition step that would preclude the competing evolution of the alkyl-Ni(I) intermediate via β-hydride elimination/reinsertion pathways, which favors the formation of the linear isomer.^21,22 This observation, together with the experimental ρ value obtained from the Hammet study, may point to the reductive elimination being rate limiting.^42,43

Scheme 5.

Proposed mechanism for the photocatalytic acylation of gaseous alkanes

In conclusion, we have developed an efficient protocol for the direct cross-coupling of acid chlorides with gaseous alkanes. The reaction proceeds under mild conditions, operates using simple reactors, and provides access to a range of industrially and commercially relevant ketones in good yields with high regioselectivity. The vast availability of the substrates and the importance of the synthesized ketones as commodity chemicals add significant value to this protocol, which represents a potential tool for the valorization of these abundant chemical feedstocks.

Methods

Materials and methods

Methane (99.5% purity), ethane (99.5% purity), propane (99.5% purity), and isobutane (99.5% purity) were purchased from Nippon Gases. All the acyl chlorides and other reagents were purchased from Aldrich Chemical, Alfa Aesar, BLD Pharmatech, Fluorochem, or TCI Europe N.V. and used without further purification, unless noted otherwise. The TBADT catalyst,⁴¹ nickel catalysts,⁴⁴ and nickel acyl complex⁴⁵ were prepared according to reported procedures. The LED light sources used are Kessil PR 160L-370-G2 (43 W) and 390 nm (52 W). Analytical thin-layer chromatography was carried out on silica-coated aluminum plates (silica gel 60 F254 Merck), and compounds were visualized using 254 nm UV light. Flash column chromatography was performed using a Buchi Pure Chromatography System with FlashPure EcoFlex Silica 4 to 80 g, unless stated otherwise. GC-MS analyses were performed in an Agilent instrument GC-8890 equipped with a chemical ionization (CI) MS-5977B detector. High-resolution MS was carried out on a Bruker microTOF spectrometer using atmospheric pressure chemical ionization-flow injection analysis (APCI-FIA). ¹H-, ¹³C-, and ¹⁹F-nuclear magnetic resonance (NMR) experiments were carried out using a Bruker AVIII-500 MHz, a Varian Mercury 300 MHz, or an Agilent VNMRS-300 MHz NMR spectrometer. Chemical shift values are reported in ppm, with the solvent resonance as the internal standard (CHCl₃: δ 7.26 for ¹H and δ 77.2 for ¹³C). Coupling constants (J) are given in Hz.

Please see the supplemental methods for synthetic procedures, detailed optimization studies, mechanistic studies, and ¹H-NMR, ¹³C-NMR, and ¹⁹F-NMR spectra.

Reaction setup

The low-pressure reactions (1.5–6 bar) were set up using Büchiglasuster tinyclave reactors equipped with a steel screw cap with 2 openings of 1/8″ NPT for Swagelok fittings with bursting discs and a manometer valve. The reaction vessel volume was 10 mL with an outer protective mesh. A single LED lamp (Kessil PR 160L-370-G2) was positioned at 1 cm. A fan was utilized to keep the temperature constant (25°C), and the stirring rate was set at 1,200 rpm (Figure S1). Reactions above 10 bar were performed in 20 mL glass vials (Pobel), which were then placed inside a high-pressure photoreactor assembled from a standard high-pressure reactor (Parr series 4760, non-stirred), composed of a base of 6.3 cm total diameter and a head (height: 6.5 cm, internal diameter: 3.0 cm). The head contains a centered sapphire window (diameter: 2.54 cm), two inlet/outlet valves for charging/discharging gases, and a safety release valve (safety pressure: 10 MPa at 200°C). A plastic 3D-printed piece was used as a holder for the LED lamp to ensure constant light distance and airflow. A single LED lamp (Kessil PR 160L-370-G2) was positioned on the holder, illuminating the reaction through the sapphire window. A fan was utilized to maintain a constant temperature (25°C), and the stirring rate was set at 1,200 rpm (Figure S2).

Resource availability

Lead contact

Further information and requests for resources should be directed to and will be fulfilled by the lead contact, Martín Fañanas-Mastral (martin.fananas@usc.es).

Materials availability

All materials may be prepared as described in the supplemental methods.

Data and code availability

• •All other data are available in the supplemental information.
• •Any other data used for analysis not included in the supplemental information are available from the lead contact upon reasonable request.
• •No original code is reported in this study.

Acknowledgments

Financial support from the European Research Council (ERC-CoG 863914-BECAME), the Agencia Estatal de Investigación (TED2021-132449B-I00), the Xunta de Galicia (ED431C 2022/27; Centro de investigación do Sistema universitario de Galicia accreditation 2023–2027, ED431G 2023/03), and the European Regional Development Fund (ERDF) is gratefully acknowledged. A.M.N. thanks the European Union’s Horizon Europe research and innovation program for a Marie Skłodowska-Curie postdoctoral fellowship (MePhoCat, no. 101150274). J.M.M. thanks Xunta de Galicia for a predoctoral fellowship (ED481A-2024-112).

Author contributions

A.M.N., S.B.-A., and J.M.M. conducted all the experiments. A.M.N. and M.F.-M. designed the experiments. A.M.N. and M.F.-M. wrote the manuscript. M.F.-M. directed the research.

Declaration of interests

The authors declare no competing interests.

Published: October 20, 2025