Synthesis of gamma-nonalactone (Coconut aroma)

This is a very easy preparation of gamma-nonalactone, or 4-pentyl-gamma-butyrolactone, a naturally occuring aroma.

The downside is that it starts from heptaldehyde, which isn't available to nayone... I gues it could be made by oxidizing heptanol by TEMPO-mediated reactions, but you still need to find the alcohol...

The reaction is composed of two steps:

-malonic synthesis of a b,gamma-unsaturated acid
-intramolecular cylization to the lactone.






Sorry, no photos this time, my camera is broken for the moment...


Preparation of the unsaturated acid


4.16g (40 mmol) of malonic acid where charged in a 250mL 3-neck RBF equipped with a condenser, an addition funnel, a thermometer and magnetic stirring.
5,6 mL (40 mmol) of heptaldehyde were added, followed by 8mL triethylamine (60 mmol). the addition of the amine caused some milkyness to appear as the malonic acid dissolved, and quickly gave place to a clear, colorless solution. The mixture was heated to 100-110°C for 1H.

The now slight yellow limpid solution was transfered to an seperating funnel, and the flask rinced with 2x20mL Et2O. An aqueous layer dropped out.
40mL of 4N HCl were then added, and the layers shaken. The slightly opaque yellow ethereal layer was then washed with 2x10mL dH2O.
40mL 1.25N NaOH were then added, the aqueous layer turning light redish, and the organic staying light yellow. The aq. was seperated, and the organic washed with 2x10mL dH2O.
The solution was washed with 2x10mL Et2O, and acidified with 40mL 4N HCl, a red/purple oily compound crashing out. The acid was extracted with 3x20mL Et2O, washed with 20mL brine, and dried over Na2SO4.

The solvent was then removed, leaving 4.7g (30,13 mmol, 75,33%) of light amber oil, single spot by TLC (Pet ether:AcOEt 6:4, Rf= 0.52).

Cyclisation to the lactone

The crude product obtained in the first step was dissolved in 30mL heptane, giving a very slightly amber limpid solution. An equal weight of Amberlyst 15 cationic resin was added, and the mixture heated to reflux for 1H with strong stirring.
The flask was then cooled, and the colorless limpid supernatant decanted into a 100mL flask. The resin was washed with 3x10mL Et2O, which were added to the heptane solution.

The solvents were removed, leaving 4.6g (29.45 mmol, 97.74%) of a pale yellow oil, with a strong coconut smell, very "natural" smell.

The lactone was thus obtained in a 73.63% yield from the aldehyde.

The Amberlyst-catlyzed reaction was very clean and effective, I am surprised at the selectivity of this catalyst, considering the extremly simple workup an dthe fact that it can be recyled numerous times.. Agreed, it is a bit expensive, but worth the effort IMHO... I think I will try it with other acid-catalyzed condensations and see how it goes!


I think this reaction is generaly applicable for substitued lactones, using various aldehydes.

Well, 1-halogenohexane could bve made from the alcohol, the grignard reagent prepared and reacted with formaldehyde to the heptanol, which could then be oxidized...

In anycase, the preparation of the aldheyde is even harder than the reaction tiself, which is a pity... Maybe other (shorter) aldheydes give a simialr smell? There is another dicyclic butyrolactone wich is said to have a coconut smell:

"Coconut decanone"

This is basicly a cyclized 4-pentylbutyrolactone with a methyl attache don the 3 position of the side chain, so maybe a straight octan chain will behave similarily? Only experimenting can tell!

3-nonenoic acid

In 250 ml 3-neck RBF were mixed:
12,5 g of malonic acid (120 mmol, M = 104,06 g/mol)
13,7 g heptaldehyde (120 mmol, M = 114,19 g/mol, should be 17,0 ml, density 0.80902 at 30 °C)
18,2 g triethylamine (180 mmol, M = 101,193 g/mol, should be 25,1 ml, density 0.7255).
(The flask was on a scale - I prefer weighing reactants instead of measuring volumes).
Liebig reflux condenser was attached, a thermoprobe of heating mantle was inserted and the third neck was stoppered (using 2-neck flask would be enough, there is no need to drop liquid reactants through the third neck using dropping funnel).
The content was refluxed for 1 hour at temperature above 100 C inside the flask (heating mantle target T was set to 105 C).
At 81 C there was already little of tiny bubbling. Reflux vs. CO2 evolution ? - but the CO2 should not evolve, it should be bound by the triethylamine into the amine carbonate or hydrogencarbonate.
At 85 C there was already a condensate dripping back into the reaction from the reflux condenser.
At 87 C there were already few drops per second dripping back to the reaction from the condenser. According the b.p. it looks like some azeotrope because b.p. triethylamine 89 C, heptanal 153 C and also some reaction water may be present too. Or it was just refluxing triethylamine and the T was few C within the tolerance of the thermoprobe and heating mantle.
At 103 C white fumes observed in the flask and the condenser (below that T there was just pure reflux without fumes). The fumes were in the flask but also in the top part of the condenser. After half an hour the white fumes observed only in the flask (but in reduced quantity) and not in the condenser anymore, in the condenser there were just clear droplets of a condensate. Maybe the decarboxyllation finished (or just decomposition of triethylamine hydrogencarbonate)?

Using smaller flask 100 ml would be better (not much bubbling either CO2 evolution), less air in the apparatus = reduction of the oxidation of the aldehyde. I used too big flask due to expected massive CO2 evolution which did not occur (or occured at slow rate).

The reaction mixture was transferred into 250 ml separatory funnel, RBF was rinsed with 3 x 20 ml diethylether, ether washings were added into the separatory funnel.

Now the amine was neutralized with HCl which requires at least 180 mmol of HCl. It is exothermic which could boil out the ether which is undesirable. Strongly cooled HCl was prepared by mixing 20 g of crushed ice with 13,0 ml of 35 % HCl.
20 g of crushed ice was added into the separatory funnel and then the above chilled diluted HCl dropwise in portions while stirring. Bottom water phase was discarded, testing with pH paper showed strong acidic reaction (if not add few more drops of diluted HCl), upper organic layer was kept.

The ether layer was washed with 2 x 20 ml of water, bottom water phase discarded and the upper organic layer kept.

Now the 3-nonenoic acid was neutralized into its sodium salt. It requires theoretically 120 mmol of NaOH (4,8 g).
5,2 g NaOH (slight excess) was dissolved in 80 ml of water + 20 g of crushed ice. Chilled solution of NaOH was added dropwise into the separatory funnel to the ethereal solution of the 3-nonenoic acid dropwise while stirring. It is again exothermic. Do not allow ether to boil out, eventually add 10 g of crushed ice into the separatory funnel.
Bottom water phase containing sodium 3-nonenoate was kept and upper ether layer discarded (ether could contain unreacted heptanal and organic sideproducts and so on, Klute extracted the upper organic layer twice with water and added it to the water phase of sodium nonenoate to increase yield, but I didn't and my yield was comparable good).
Water phase was then extracted with 2 x 20 ml of ether (to remove e.g. unreacted heptanal and possible organic sideproducts). The upper ether phases discarded and bottom water phase was kept.
Water solution of sodium 3-nonenoate was returned into the separatory funnel. Testing with pH paper showed strongly alkaline reaction pH above 12. If there were no significant loses of the product in the previous extraction step, it should prove that the product is not mainly dicarboxylic acid and that its decarboxylation occurred.

Now there is necessary convert sodium salt back into free acid which requires 120 mmol of chilled HCl = 8,5 ml of 35% HCl plus some excess to neutralize the small excess of NaOH.
120 mmol of HCl = 8,5 ml of 35% HCl.
To the separatory funnel was added cold 11 ml 35% HCl dropwise while stirring (slight excess due to previous slight excess of NaOH). Exothermic! 9 ml 35% HCl was not enough - tested with pH paper (just week acidic pH and still not strongly acidic), but after 11 ml HCl the pH paper showed highly acidic result. There was some ether present (estimated 10-15 ml) - slightly larger volume of the upper organic layer observed than a volume of only sole 3-nonenoic acid. After cooled down 10 ml more ether was added, shaken and separated. The upper organic phase containing 3-nonenoic acid was kept and bottom water phase was extracted with 2 x 20 ml ether. The ether extracts were combined together and washed with 50 ml of concentrated NaCl solution.

Then the organic layer was transferred into 100 ml Erlenmeyer flask and dried with anhydrous Na2SO4.

Diethylether was distilled out.

16,9 g of pale yellow product obtained (108 mmol = 90 %, M = 156.22 g/mol)
COOL !!! This is more than the weight of the reactant heptanal also its scent is completely different from the scent of the heptanal and very close to the scent of nonanoic acid.



photos:



reagents





reactants mixed in RBF





small bubbling already at 81 C





T increases





at 103 C white fumes observed in the flask and also at the top of condenser





After half an hour the fumes only in the RBF and not in the condenser anymore





After 1 hour of reflux around 105 C the reaction was done.





After addition of diethylether and acidification with HCl. Upper organic layer kept. Bottom water phase discarded after checking that its pH is strongly acidic.





After neutralization with NaOH. Bottom water phase containing sodium 3-nonenoate kept (pH strongly alkaline due to small excess of NaOH) and upper ether phase discarded (unreacted heptanal + organic sideproducts and so on).





After converting sodium salt to free acid using HCl. Upper ether layer kept and bottom water phase discarded (pH of water phase strongly acidic on universal indicator paper, using small excess of HCl). Note this photo is soon after vigorous shaking before both phases separated sharply.





Washing ether extract with saturated water solution of NaCl to reduce water present in the ether. Upper ether layer kept and bottom water phase discarded.





drying with Na2SO4





distilling out ether





3-nonenoic acid, the desired product

gamma nonalactone

84% H2SO4 was prepared this way:
4,0 g H2O + 30,0 g 95% H2SO4 (28,5 g 100% H2SO4) = 28,5/34 = 84%
To a beaker with 4,0 g H2O was dropwise added 30,0 g 95% H2SO4 while stirring and cooling the beaker on ice-water bath (extremely exothermic !!!)

To 16,9 g of 3-nonenoic acid in 100 ml RBF was added dropwise while stirring the above 34 g of cooled 84% H2SO4. The both liquids were completely miscible (seems H2SO4 immediately adds to the C=C double bond ?). Stir bar added and the reaction was refluxed while magnetically stirring (maybe the stirring is unnecessary, but I have good experience in preventing bumping using magnetic stirring). I suppose the reflux condensate is an azeotrope of H2O + 4-pentyl-gamma-butyrolactone as there were a mixture of water with immiscible liquid visible in a form of tiny white droplets in the condenser. The reaction was refluxed for 1 hour (maybe shorter time or lower temperature are enough and better ???). External infrared thermometer showed T inside 98-107 C (the accuracy of the thermoprobe is very likely something like +-5 C). Rancid gas evolved (shorter chain organic acids or HCl from remainders of NaCl ???) so an adapter was inserted into the top of the condenser to which a hose was connected and its open end was led outside of lab.

30 g 95% H2SO4 (M = 98 g/mol) means (30 * 0,95) / 98 = 0,29 mol H2SO4.
NaHCO3 molar mass 84 g/mol. 0,58 mol NaHCO3 = 48,7 g.

50g NaHCO3 (0,60 mol, slight excess) + 250 ml H2O were mixed in 500 ml Erlenmeyer + glass stirring rod inserted. NaHCO3 is not fully soluble in the above amount of water but it is going to be reacted with H2SO4. The environment needs to be as close to neutral as possible. More alkaline as well more acidic environment would hydrolyze the lactone.

To the Erlenmeyer flask was dropwise added in portions while stirring the reaction mixture (cca not more than 0,5 ml evey time). The process is very slow and messy due to bubbling+foaming and perhaps some soap like sodium salt of organic acid present. At the end also RBF was rinsed with the contenet from the Erlenmayer flask and the content was returned to the Erlenmeyer flask (but a lot of tar stayed adhering to RBF). More than 3 hours of workup was required and the process was very messy as the content in the RBF was very thick, contained a lot of tar. Dirtyness and slowness is comparable with workup of phenol nitration. Use Klute method rather (clean, fast, high yielding).

The content was transferred into 500 ml separatory funnel and the product extracted with 3 x 30 ml of diethyl ether.

Ether extracts were combined and washed with 50 ml of saturated water solution of NaCl.

Organic layer dried with Na2SO4.

Ether distilled out.

A little of very impure product was obtained which was unsuitable for vacuum distillation. Steam distillation was rather tried to test whether the lactone does not easily hydrolyze. Clevenger apparatus was used. Do not just steam distill without such apparatus. The solubility of lactone in water is slightly more than 1 g per liter. You need Clevenger apparatus so the water returns back into the distilling flask.
A little of red-brown product obtained.
Its scent was noticeable already during workout. I perceived it like 50% coconut + 50% fruit (closest to peach). It is not pure coconut but also a lot of fruit (mostly peach). If you would like pure peach, start not from heptanal (intermediate 3-nonenoic acid) but from octanal (intermediate 3-decenoic acid and product gamma decalactone) - read the following post containing messy compilation of various useful data gathered from internet. Heptanal could be obtained by distilling castor oil with colofony pine resin at elevated temperature while applying water pump vacuum.



photos



mixed 3-nonenoic acid + 84% H2SO4





heating + stirring the reaction mixture already darkened





the reaction much more darkened during time





there were small droplets of immiscible organic compound visible in the condenser (careful eye required, do not confuse with plenty of air bubbles in cooling liquid)





completely dark black and tarry mixture obtained





long lasting and messy neutralization with NaHCO3




ether extraction, the layers almost impossible to distinguish, a small hand lamp helped, this is the first extraction, the second extraction is slightly better visible and the third extraction already well visible, here only the first extraction before and after draining out some bottom water phase





ether extractions combined and dried with Na2SO4





ether distilled out





crude product





purification by steam distillation and Clevenger apparatus



[Edited on 28-8-2024 by Fery]