Maria do S. B. da Silva,
Jhudson G. L. de Araujo,
Júlia C. C. V. Bento,
Amanda M. de Azevedo,
Carlos R. O. Souto,
Aécia S. D. dos Anjos,
Aruzza M. M. de Araújo,
Djalma R. da Silva,
Fabrício G. Menezes,
Amanda D. Gondim and
Lívia N. Cavalcanti*
Federal University of Rio Grande do Norte, Institute of Chemistry, 59072-970, Natal, RN, Brazil. E-mail: livia.cavalcanti@ufrn.br
First published on 29th September 2022
An operationally simple and highly selective method for the decarboxylation of fatty acids under remarkably mild conditions is described herein. The activation of the aliphatic carboxylic acids by esterification with N-hydroxyphthalimide (NHPI) enabled efficient deoxygenation to synthesize n-alkanes in up to 67% yield, employing inexpensive PMHS as a hydrogen source, NiCl2·6H2O, bipyridine, and zinc in THF. In contrast to the conventional thermo-catalytic approaches, this protocol does not require high temperature and high pressure of hydrogen gas to deoxygenate biomass-derived carboxylic acids, thus representing an attractive alternative for producing drop-in biofuels.
Vegetable oils are one of the most prominent types of renewable feedstock to produce biofuels such as biodiesel.4 However, as well as the crude oils, the fatty esters are still associated with poor fuel properties such as low heating value, high viscosity, and low oxidative stability.5,6 Additionally, the high oxygen content of this biofuel also results in poor cold flow properties, strongly limiting their use as blends with fossil fuels for diverse applications.7 Therefore, technical approaches for deoxygenating vegetable oils and their derived biofuels are important targets for increasing their compatibility with petroleum-based fuel infrastructure.
Several protocols have demonstrated the utility of vegetable oils for the production of biofuels through deoxygenation pathways.8–11 These reactions enable the conversion of biobased triglycerides, fatty acids, or esters into mixtures of renewable hydrocarbons that are able to either partially or completely replace fossil fuels as they are functionally equivalent to them.12 Remarkably, among the various types of suitable substrates for deoxygenation, fatty acids obtained upon hydrolysis of vegetable oils have met biofuel application more promisingly due to their low cost, availability, and carbon number lying within the range of fossil hydrocarbons.13,14
Depending on the biomass employed as feedstock, many technological routes such as hydroprocessing, catalytic hydrothermolysis, alcohol-to-jet, and gasification combined with the Fischer–Tropsch process are available for accomplishing the synthesis of renewable hydrocarbons.15,16 Particularly for the conversion of fatty acids and their derivatives, hydroprocessing has been a widely exploited approach17–19 (Scheme 1A). Although hydroprocessing has become increasingly relevant in petrochemical industry settings,20 it demands high temperatures (>250 °C) and high pressure (>10 bar) of hydrogen gas, which is usually obtained from non-renewable sources.21,22 Thermo-catalytic pyrolysis is a useful alternative that allows conversion without hydrogen. However, this process lacks selectivity, resulting in mixtures of products that require further upgrading to provide suitable biofuels.23–25 Hence, efficient deoxygenation methods that are operational in low temperature and ambient pressure under a hydrogen-free atmosphere are highly desirable.
In the past few years, radical decarboxylation of redox-active esters (RAE) using metal catalysts under mild conditions has become a promising approach for forging C–C, C–X, and C–H bonds, introducing complexity in a variety of molecules with remarkable selectivity.26,27 Recently, Baran and coworkers reported the use of a Ni-based catalyst for the decarboxylation of RAEs in the presence of phenylsilane or Michael acceptors to afford hydrocarbons or coupling products respectively28 (Scheme 1B).
Although this protocol proved to be efficient to synthesize a variety of useful products from several examples of structurally diverse carboxylic acids, fatty acids were not included in the scope of the reductive method for hydrocarbon synthesis. Based on this important work, we envisioned the production of renewable hydrocarbons from bioderived fatty acids by reductive decarboxylation of RAEs under mild conditions using a Ni catalyst and a silane as the hydrogen source (Scheme 1C).
Initial optimization studies revealed, after extensive solvent screening (see ESI for full details†), that THF was the ideal solvent for the transformation, resulting in the conversion of RAE into the desired hydrocarbon in 45% yield along with trace amounts of radical homocoupling byproduct (see Table ES1 in ESI†). Remarkably, the use of a mixture of THF:DMF:iPrOH (10:2:1) (Table 1, entry 2), which was advantageous for the decarboxylation of a range of RAEs reported by Baran and coworkers,28 decreased the yield of 2a to 22%. Other Ni-based catalysts such as NiBr2 and Ni(OAc)2·4H2O (Table 1, entries 3 and 4) as well as other ligands such as dtbbpy and DPPB (Table 1, entries 5 and 6) also proved to be less effective to afford the decarboxylation product. Increasing the stoichiometry of both Ni and ligand diminished the yield of 2a (Table 1, entry 9).
Entry | Modification from standard conditions | Yieldb (%) | |
---|---|---|---|
2a | 3a | ||
a Reactions were run on a 0.1 mmol scale.b Yield determined by GC-MS using hexadecane as internal standard. PMHS: poly(methylhydrosiloxane); bipy: 2,2-bipyridine; dtbbpy: 4,4-di-tert-butyl-2,2-dipyridyl; DPPB: 1,4-bis(diphenylphosphino)butane; PhSiH3: phenylsilane; TMDSO: 1,1,3,3-tetramethyldisiloxane. See ESI for additional details. | |||
1 | None | 45 | 2 |
2 | THF:DMF:iPrOH (10:2:1) | 22 | 0 |
3 | NiBr2 as catalyst | 19 | 0 |
4 | Ni(OAc)2·4H2O as catalyst | 10 | 2 |
5 | Dtbbpy as ligand | 28 | 0 |
6 | DPPB as ligand | 6 | 0 |
7 | PhSiH3 as hydrogen source | 28 | 26 |
8 | TMDSO as hydrogen source | 0 | 0 |
9 | 20 mol% NiCl2·6H2O + 40 mol% bipy | 6 | 0 |
10 | Mn as reducing agent | 0 | 0 |
11 | Addition 3 equiv. K2CO3 | 46 | 7 |
12 | Reaction concentration of 0.2 M (0.5 mL) | 36 | 0 |
13 | Reaction at 70 °C | 35 | 8 |
14 | No nickel or ligand | 0 | 0 |
15 | No PMHS | 0 | 0 |
16 | Methyl laureate or lauric acid instead of NHPI ester | 0 | 0 |
Replacing PMHS with different hydrogen sources commonly reported for this type of transformation39,40 significantly decreased the formation of the desired product under this set of conditions (Table 1, entries 7 and 8). Notably, other reductive agents such as Mn (Table 1, entry 10) completely inhibited the decarboxylation. The presence of a base such as K2CO3 as an additive (Table 1, entry 11) did not improve the yield of 2a, favoring radical homocoupling instead. Furthermore, increasing either reactants concentration or reaction temperature to 70 °C was also prejudicial to the formation of 2a (Table 1, entries 12 and 13). Control studies confirmed the necessity of Ni catalyst, ligand, and hydrogen source for the reaction to occur (Table 1, entries 14 and 15). Additionally, no product of decarboxylation was observed when using either the methyl laureate or lauric acid as substrates, proving that NHPI esters are crucial reactants to enable efficient reactivity (Table 1, entry 16).
After optimization experiments, the efficiency of the reductive deoxygenation of 1a using PMHS (2 equiv.), NiCl2·6H2O, bipyridine, and Zn in THF at 40 °C was confirmed by GC-FID with a calibration curve, which allowed to determine that 2a could be synthesized in 51% yield. It is noteworthy that, under the applied conditions, the reaction exclusively afforded hydrocarbons as products. Both undecane (n-C11) and docosane (n-C22) are valuable biofuel candidates since they have carbon number in the jet-fuel and diesel range, respectively. Furthermore, the N-phthalimide formed during the reaction work up, can be recovered and utilized to synthesize more N-hydroxyphthalimide41 that can be converted into 1a upon reaction with lauric acid.
With suitable conditions in hand, the generality of the decarboxylation method was evaluated. The scope of the fatty acids-derived RAE was examined by employing substrates with different carbon chain lengths. Reaction yield was determined by GC-FID using a calibration curve. The devised protocol proved to be effective for converting RAE 1b, which was synthesized from capric acid, into n-nonane in 45% yield, and octadecane in 2.5% yield (Table 2, entry 1). A better result was observed for the synthesis of n-tridecane (65% yield) after decarboxylation of 1c, which also resulted in the formation of the homocoupling product in a 4.9% yield (Table 2, entry 2). The reaction of the RAEs derived from palmitic and stearic acids delivered n-pentadecane in a 67% yield, and n-heptadecane in a 54% yield while the formation of the respective byproducts was diminished to less than 5% (Table 2, entries 3 and 4). Extending the scope to RAE 1f resulted in the complete suppression of 3f generation along with lower yields of 2f in relation to the other synthesized hydrocarbons. It is possible that the formation of 2f was limited by the reaction between the radicals formed upon decarboxylation and the unsaturation within the substrate, leading to a polymerization product that could not be detected by the employed technique.14
Entry | Feedstock | Principal product | Yieldb (%) | |
---|---|---|---|---|
2a | 3a | |||
a General decarboxylation reaction conditions: 0.1 mmol of redox-active ester, 0.2 mmol of PMHS, 0.01 mmol of NiCl2·6H2O, 0.02 mmol of bipyridine, 0.05 mmol of zinc in 1 mL of THF (0.1 M).b Yield determined by GC-FID using calibration curve. NPhth: N-phthalimide. | ||||
1 | n-C9 2b | 45 | 2.5 | |
2 | n-C13 2c | 65 | 4.9 | |
3 | n-C15 2d | 67 | 4.2 | |
4 | n-C17 2e | 54 | 1.7 | |
5 | n-C17:1 2f | 33 | 0 |
Although moderate yields were determined for the n-alkanes, all reactions proceeded with more than 98% conversion, except for ester 1f, which afforded the respective product with an 85% conversion. These data indicate that the redox-active esters are likely to decompose under the reaction conditions, which is corroborated by the presence of fatty acids and phthalimide in crude reaction mixtures according to GC-FID analysis.
The optimized conditions were also employed to convert the fatty acids using a one-pot reaction approach. The decarboxylation protocol was applied to RAEs generated in situ to produce n-alkanes. Notably, the RAEs from lauric and palmitic acids (Table 3, entries 1 and 4) delivered the corresponding hydrocarbons in similar yields in relation to the protocol that starts from the isolated esters. For the other substrates, however, significantly lower yields were observed, especially in the decarboxylation of the RAE from oleic acid. Similar outcomes were already described for this type of transformation.28,42,43
Entry | Feedstock | Principal product | Yieldb (%) | |
---|---|---|---|---|
2a | 3a | |||
a Activation in situ conditions: fatty acid (0.1 mmol), N-hydroxyphthalimide (0.11 mmol), DIC (0.11 mmol) in DCM at rt, then 0.2 mmol of PMHS, 0.01 mmol of NiCl2·6H2O, 0.02 mmol of bipyridine, 0.05 mmol of Zn in 1 mL of THF (0.1 M).b Yield determined by GC-FID with calibration curve. NPhth: N-phthalimide. | ||||
1 | n-C11 2a | 48 | 4.8 | |
2 | n-C9 2b | 28 | 1.2 | |
3 | n-C13 2c | 27 | <1 | |
4 | n-C15 2d | 51 | 2.3 | |
5 | n-C17 2e | 24 | 1.7 | |
6 | n-C17:1 2f | 10 | 0 |
Finally, the developed method was applied to a mixture of RAEs synthesized from saturated fatty acids. Fortunately, the materials could be converted into a mixture of hydrocarbons (C9–C17) within the diesel and jet-fuel ranges in 67% yield (Fig. 1). This result demonstrates the potential of the decarboxylation methodology to produce n-alkanes compatible with drop-in biofuels from more complex matrixes such as mixtures of fatty acids originating from vegetable oils or animal fat.
Fig. 1 Chromatograms of the decarboxylation reaction of REAs mixture derived from fatty acids. (A) Initial reaction sample; (B) final reaction sample; Sd: internal standard. |
Based on these results, the proposed mechanism is analogous to previously reported Ni-catalyzed reductive decarboxylation of RAE (Scheme 2).27,44–46 On the basis of this mechanism the BiPy ligand is necessary to help stabilize the in situ formed Ni(I) catalyst II that undergoes transmetallation with PMHS to afford intermediate III. This complex would then donate an electron via SET to the RAE forming V and the radical anion IV which undergoes radical fragmentation (decarboxylation) yielding the primary alkyl radical VII and anionic phthalimide VI. Addition of radicals VII and VI to intermediate V gives intermediate VIII which after reductive elimination affords the desired alkane product and regenerates the catalyst.
Footnote |
† Electronic supplementary information (ESI) available. See https://doi.org/10.1039/d2ra04057c |
This journal is © The Royal Society of Chemistry 2022 |