Open Access Article
Arthur
Lasbleiz
,
Pierre-Alexandre
Deyris
,
Franck
Pelissier
,
Yves-Marie
Legrand
,
Claude
Grison
* and
Claire M.
Grison
Bio-inspired Chemistry and Ecological Innovations (ChimEco), UMR 5021 CNRS-University of Montpellier, Cap Delta, 1682 rue de la Valsière, 34790 Grabels, France. E-mail: claude.grison@cnrs.fr
First published on 31st October 2025
We described the first synthesis of biobased ethyl and methyl formates. The synthetic strategy is based on the transesterification of natural geranyl and citronellyl formates derived from Pelargonium species essential oils with ethanol or methanol, promoted by an ecocatalyst®. Good yields (up to 80%) were obtained with an excellent selectivity.
Sustainability spotlightThe work described in this article goes beyond the boundaries of sustainable chemistry: first, it adheres to all the principles of green chemistry described by Anastas and Warner. This study quantitatively assesses, through life-cycle analysis (LCA), the possible environmental benefits of replacing organometallic complexes (Ru, Pd, Co, Cu–Zn, or Fe-based catalysts) with biosourced catalysts, called ecocatalysts. Ecocatalysts are natural, non-toxic, non-ecotoxic, efficient and abundant. Elimination of hazardous substances in production and use is demonstrated. No synthetic or petroleum-based inputs are used. The process is simple and straightforward. It generates no waste and improves the quality of the raw material, Pelargonium graveolens essential oil. Second, the proposed strategy involves massive harvesting of the world's most invasive plant species, exotic knotweeds. Using the cut parts of these plants contributes to management efforts for these plant species by weakening them. Their use in organic synthesis also supports the repeated harvesting of these plants. In other words, this article shows how sustainable chemistry can contribute to ecological solutions aimed at controlling the development of invasive plant species. This is an additional principle that can inspire future research in sustainable chemistry. |
Different methods for the preparation of EF and MF are described in the literature. Industrially, they are derived from the carbonylation of ethanol or methanol.33 However, this carbonylation requires high-purity CO gas,34 which is difficult to obtain without any trace of CO2, which is an inhibitor of ruthenium-based catalysts used in the reaction.
At the laboratory scale, these two esters have usually been prepared by esterification of formic acid.35 Nowadays, many research articles are devoted to the reduction of CO2 to small organic molecules. The O-formylation of methanol or ethanol using CO2 and Ru, Pd, Co, Cu–Zn or Fe-based catalysts has been reported36–38 but in each case, the yields remained moderate. Cannon et al. showed the importance of the pH effect in these reactions.39 Unlike the above syntheses performed in basic medium, the hydrogenation of CO2 in acidic medium, catalysed by phosphine-ligated Ru catalysts, favours the formation of MF.39 Electrochemical reduction of CO2 in acidic ethanol on Pb and Sn cathodes also leads to the production of EF. In this reaction, the formed formic acid serves as an autocatalyst to perform its own in situ Fischer esterification with ethanol and therefore produces EF with an interesting yield.40 Promising results were obtained with the Ni-catalyzed hydrosilylation of CO2 followed by an O-formylation of alcohols in the presence of HBF4.41 The benefit of these acidic conditions was also described when using boryl formate in the TsOH-catalysed O-formylation of benzylic alcohols.42 Similarly, the presence of triformatoborohydride, which is formed during the reaction of NaBH4 with CO2,43 could act as an O-formylation reagent with formic acid as a catalyst.44 These acidic conditions suggest that the O-formylation step proceeds via a transesterification reaction.40
All these new approaches to the synthesis of formates are driven by the efforts to decarbonize industrial activity coupled with the valorisation of CO2, which is largely responsible for the greenhouse effect. However, these approaches suffer from the use of expensive, toxic and/or ecotoxic complex metal catalysts. Additionally, hydrosilane or borohydride reducing agents lead to the formation of problematic wastes regulated by REACH.
In this article, we propose to develop the first fully biobased synthesis of EF and MF using essential oils (EOs) derived from Pelargonium species, which are rich in geranyl and citronellyl formates (GF and CF, respectively) and an ecocatalyst® acting as a biosourced base catalyst, in accordance with an eco-responsible approach.
In this article, we propose to use the natural abundance of CF and GF contained in Pelargonium graveolens L’Hér. as a natural raw material to prepare EF and MF.
The synthetic strategy is based on the transesterification of these two formates with ethanol or methanol. The reaction is promoted by an ecocatalyst® (Eco1-E-FA-JA), a mineral catalyst of 100% plant origin. The selected plant species is Fallopia japonica, an invasive alien species (IAS) of wetlands. Its intensive harvesting helps weaken it and control its proliferation by exhaustion (Fig. 1).
| Element | Al | Ca | Fe | K | Mg | Na |
|---|---|---|---|---|---|---|
| wt% rsd | 0.10 ± 0.00 | 6.78 ± 0.08 | 0.15 ± 0.02 | 30.75 ± 0.13 | 3.64 ± 0.02 | 0.15 ± 0.00 |
The Mg content is 3.64 wt%, which is high in comparison to classical plants, but reasonable given the high chlorophyll content in the aerial parts. The K content (30.75 wt%) is remarkably high but falls within the main specificities of many IAS.
Ion chromatography analyses give the concentration of the anions from the ecocatalyst® soluble in water. Phosphate PO43−, sulfate SO42−, carbonate CO32− and chloride Cl− anions were detected in significant quantities (Table 3 and SI part 1.4 & 2.2).
| Anion | PO43− | SO42− | CO32− | Cl− |
|---|---|---|---|---|
| wt% (g of anion/100 g ecocatalyst) ± rsd | 1.90 ± 0.02 | 7.20 ± 0.11 | 11.10 ± 0.10 | 5.40 ± 0.09 |
The amount of the analysed sulfates is particularly high (4 times higher than phosphates). Chlorides and carbonates are also formed in significant quantities. Given the high carbonate content, the pH of the solution was measured and found to be 11.9 (200 mg of the ecocatalyst® in 10 mL of water). This basicity is higher than the pKa2 of a simple carbonate, which makes the ecocatalyst® interesting and appropriate to be used in organic synthesis.
X-Ray Powder Diffraction (XRPD) studies were carried out to determine the nature of the combinations between the different mineral elements in the crystalline state (Table 4 and SI part 1.5 & 2.1).
| Name | Formula |
|---|---|
| Sylvite | KCl |
| Potassium calcium phosphate | KCaPO4 |
| Potassium sulfate | K2SO4 |
| Potassium carbonate sesquihydrate | K2CO3(H2O)1,5 |
| Fairchildite | K2Ca(CO3)2 |
| Calcite | CaCO3 |
| Periclase | MgO |
During heat treatment, the self-assembly of the mineral elements led to the formation of two unusual but interesting salts, which are rarely used in organic synthesis:
- a double carbonate of potassium and calcium, fairchildite, K2Ca(CO3)2
- a mixed phosphate of potassium and calcium, KCaPO4.
The presence of carbonates, and fairchildite (K2Ca(CO3)2) in particular, was expected in Eco1-E-FA-JA, as it has been found previously in other ecocatalysts® derived from F. japonica.50 Likewise, the presence of mixed phosphates was consistent with previous ecocatalysts® derived from F. japonica.50 However, the phosphate species differed from only one cation from previous work, as KCaPO4 was observed here, instead of KMgPO4. This slight difference might be attributed to the harvesting period of F. japonica and hence its physiological composition.
Moreover, another crystalline species with basic properties was detected: MgO, despite the modest temperature of the heat treatment. The presence of MgO is advantageous since this salt brings a dual character to the ecocatalyst®: the presence of oxide confers Brønsted basic properties while the magnesium shows Lewis acid properties.51
Regarding these analyses, Eco1-E-FA-JA exhibits advantageous properties to act as an excellent catalyst for transesterification reactions.
Taking as a reference the total quantity of GF and CF present in the native Pelargonium graveolens EO, 10 eq. of ethanol were used. A detailed study of the impact of the catalytic charge on CF and GF conversion rates was carried out over 28 h (Fig. 2). While the reaction without a catalyst only gave 10% conversion of CF and GF, the reaction was studied between 0.04 and 0.21 equivalents of carbonates relative to both citronellyl and geranyl formates. According to the ion chromatography results, the ecocatalyst comprises 42% soluble carbonates, and therefore, 42% of the potassium salts are in the form of carbonates. Thus, 0.5 equivalents of potassium correspond to 0.21 equivalents of potassium carbonates (including fairchildite and potassium carbonate).
![]() | ||
| Fig. 2 Effects of different Eco-1-E-FA-JA loads (0.04 to 0.21 eq. soluble carbonates) on the conversion of citronellyl (a) and geranyl (b) formates versus time. | ||
The monitoring was performed by GC-FID using external quantification of the two esters (see SI, part 1.3). Comparing the curves depicted in Fig. 2(a and b), the conversion times for the same catalytic load follow the same trend, suggesting that GF and CF exhibit a comparable reactivity.
The catalytic loading had a significant impact on the kinetics of the reaction. Indeed, using 0.04 or 0.08 eq. CO32− did not allow the reaction to reach full conversion after 28 h. While 0.13 and 0.17 eq. CO32− provided 98% conversion of both formates in 24 and 20 h, respectively, 0.21 eq. CO32− allowed the same conversion to be reached in 8 h. It is also important to note that the other terpenes contained in the EO of P. graveolens were not affected by the reaction and no byproducts were observed, highlighting the selectivity of the reaction at the same time (see SI, part 3.2 and Fig. 2).
The efficiency of the reaction was evaluated by recovering the formed EF by distillation of the reaction medium. Indeed, its boiling point at atmospheric pressure is 54 °C, which allows its separation from the EO and solvent by fractional distillation. This process yielded EF at 80% as a colourless liquid, and the excess of ethanol was recovered by increasing the bath temperature. The difference between the optimum yield and the conversion of the substrates is explained by losses of materials during distillation, linked to the fact that the product has a very low boiling point. However, this result remains very satisfactory and shows that the transesterification of the formates naturally present in the EO is a good method for producing EF. The remaining EO collected in the distillation residue was separated from the catalyst by simple filtration. Its composition was evaluated by GC-FID and compared to the starting EO (Fig. 3).
A comparison of the activity of the ecocatalyst Eco1-E-FA-JA with common basic catalysts and reagents was carried out (Table 5). This comparative study was carried out under the best conditions obtained with the ecocatalyst (0.21 equivalents of soluble carbonates (K2Ca(CO3)2 + K2CO3), 6 h at ethanol reflux). Eco1-E-FA-JA was first compared to KOH, a reagent classically used in transesterification reactions. Initially, 0.21 equivalents of KOH were used to mimic the 0.21 equivalents of soluble carbonates present in Eco1-E-FA-JA (see ion chromatography analyses, Table 3). The conversion rates of citronellyl and geranyl formates were low (46 and 30%, respectively) instead of 90 and 90% with Eco1-E-FA-JA. By introducing excess KOH (0.87 equiv. then 1 equiv.), the conversion rates of citronellyl and geranyl formates reached 88–89% and then 95–97%. However, treatment of the aqueous phase (see SI Section 3.3) revealed the formation of potassium formate (23 and 91%, respectively, see SI Fig. 14 and 15), resulting from the hydrolysis of citronellyl and geranyl formates. The ecocatalyst Eco1-E-FA-JA is therefore much more advantageous than KOH; it is both more reactive (90% conversion rate) and does not lead to any secondary hydrolysis reactions.
| Catalyst or reagent | Formate equivalent | Citronellyl formate conversion (%) | Geranyl formate conversion (%) |
|---|---|---|---|
| a 0.21 equiv. K2CO3 + 0.16 equiv. MgO + 0.43 equiv. CaSO4 + 0.17 equiv. KCl + 0.02 equiv. Ca10(PO4)6(OH)2. b 0.21 equivalents of soluble carbonates. | |||
| KOH | 0.21 | 46 | 30 |
| KOH | 0.87 | 89 | 88 |
| KOH | 1 | 97 | 95 |
| CaSO4 | 0.21 | 33 | 17 |
| CaCO3 | 0.21 | 2 | 19 |
| MgO | 0.21 | 47 | 29 |
| MgO | 1 | 88 | 72 |
| Ca10(PO4)6(OH)2 | 0.21 | 2 | 20 |
| K2CO3 | 0.21 | 83 | 77 |
| K2CO3 | 0.44 | 84 | 84 |
| K2CO3 | 1 | 100 | 92 |
| Synthetic mixture | 0.21a | 86 | 82 |
| Eco1-E-FA-JA | 0.21b | 90 | 90 |
Each mineral species of Eco1-E-FA-JA (MgO, CaCO3, CaSO4, K2CO3, K2Ca(CO3)2, KCaPO4) and a synthetic mixture of these salts were tested under the best conditions obtained with the ecocatalyst (6 h at ethanol reflux, Table 5). K2CO3 and Ca10(PO4)6(OH)2 were used to mimic K2Ca(CO3)2 and KCaPO4, respectively, because the two salts are unavailable and cannot be synthesized. As expected, conversion rates of citronellyl formate and geranyl formate were very low with CaSO4 (33–17%) and the three poorly soluble bases in EtOH, CaCO3 (19–2%), MgO (42–49%) and Ca10(PO4)6(OH)2 (2–20%). Only K2CO3 led to satisfactory conversion rates (83–77%). However, they remained slightly lower than those of the ecocatalyst. It was necessary to introduce 1 equivalent of K2CO3 to exceed the efficiency of 0.21 equivalents of soluble carbonates in Eco1-E-FA-JA (100–92% instead of 90–90%).
Finally, a synthetic mixture reconstituted from the constituent salts (0.21 equiv. K2CO3 + 0.16 equiv. MgO + 0.43 equiv. CaSO4 + 0.17 equiv. KCl + 0.02 equiv. Ca10(PO4)6(OH)2) has been studied. This composition corresponds to the results of ionic chromatography for each soluble salt (sulfate, chloride, phosphate, carbonate) and the MP-AES analysis for MgO. The conversion rates (86–82%) were slightly higher than those of K2CO3 (83–77%) and slightly lower than those of the ecocatalyst (90–90%). We can therefore conclude that the ecocatalyst Eco1-E-FA-JA is a very good transesterification catalyst. It is more active than simple petroleum-based carbonates. Its natural composition in mixed salts reinforces its reactivity.
A comparison with a catalyst frequently used in ester transesterification reactions, Ti(OiPr)4, was also performed.52 The reaction of transesterification of Pelargonium graveolens EO with EtOH and 4 mol% of this catalyst was set up in order to reproduce the conditions described by Corsi et al.52 After 8 h of reaction at reflux of EtOH, the conversions of CF and GF were measured at 42 and 58%, respectively. These conversions were far lower than the proposed ecocatalysts. The GC-FID analyses are added in part 3.4 of the SI.
Compared to the initial EO, the chemical composition of the remaining EO is significantly enriched in citronellol (40% instead of 28%) and geraniol (25% instead of 18%). This result is of significant interest, because these two terpenes are used in various applications as repellents for arthropods that carry infectious diseases,53 in fragrances54,55 and as key intermediates in organic synthesis.56,57 Moreover, the pH of the remaining EO was measured at 7, while the value of the initial EO was assessed at 2.7. The ecocatalyst® is thus able to perform quantitative transesterification reactions and also to neutralise the acidity of the initial EO in order to obtain a citronellol- and geraniol-enriched EO more suitable for human skin applications. These very encouraging results led us to extend the principle of this reaction to the transesterification of formates present in the P. graveolens EO with methanol (Scheme 2). More reactive than ethanol, methanol allows a total conversion in 4 hours.
Considering the gap between MF and methanol boiling points (32 °C and 65 °C, respectively), the distillation step took place at the same time as the reaction. While MF was directly distilled after its formation, the equilibrium of the reaction was shifted, providing a drastic reduction of the reaction duration. MF was successfully obtained with 75% yield as a colourless liquid. The excess methanol was then distilled in a subsequent fraction to be recycled. The same advantages as previously can be noted:
- Enrichment of the EO in citronellol and geraniol (SI, part 3.2 and Fig. 12 and 13).
- Neutralization of the acidity of the EO.
- Other compounds stayed untouched, and no byproducts were observed.
- Highly efficient synthesis of biobased MF and EF.
- Production of a Pelargonium graveolens EO with higher citronellol and geraniol contents.
- Neutralisation of EO acidity.
The current international market of P. graveolens EO is substantial, with nearly 300 tons produced per year. The projection of the ecocatalytic process application to this annual production could lead to form more than 18 tons of biobased EF or MF.
| This journal is © The Royal Society of Chemistry 2026 |