Vanillin

Of six solvents tried on cured vanilla pods, the polar ones, which include ethanol, extracted the most vanillin by both methods tested. Extraction was slow: at 95 °C with 66.67 mL of solvent per gram of pod, Soxhlet extraction released around 180 ppm of vanillin in 8 h, and raising the temperature from 90 to 100 °C increased the extraction by 30 per cent.

In one batch of Hainan vanilla followed through curing, glucovanillin fell from 10.67 per cent of dry weight in the fresh bean to 6.28 per cent after blanching, 0.74 per cent after sweating, 0.20 per cent after drying and 0.17 per cent in the cured bean. Vanillin moved the other way: 0.03 per cent in the green bean, 0.05 per cent after blanching, 2.30 per cent after sweating and 2.69 per cent after drying, with the highest value in the conditioned bean. Most of the vanillin therefore appeared during the six days of sweating. The vanillin gained was much smaller than the glucovanillin lost, which the authors attribute to further oxidation or reduction of vanillin to other compounds; vanillyl alcohol, absent earlier, stood at 0.22 per cent in the dried bean and 0.16 per cent after conditioning.

Comparing pods left to senesce, pods frozen and thawed, and pods cured traditionally, the authors found that traditionally cured beans reached only a medium yield of glucovanillin hydrolysis, under 50 per cent, with no measurable β-glucosidase activity and only partial breakdown of cell structure. They conclude that hydrolysis is governed by the breakdown of the compartments that keep enzyme and glucoside apart, that the level of β-glucosidase activity is not the limiting factor, and that where the compartments break down completely hydrolysis runs to completion even though most of the enzyme activity has been lost. Enzyme level, on their account, affects only the speed.

Through ninety days of traditional curing in Mexico, 23 odour-active compounds were followed in both free and bound form. Fifteen were derived from the shikimate pathway and thirteen of those were present as glucosides. The glucosides did not behave alike: most of the glucovanillin was hydrolysed in the first stages of curing, others were released only in the later stages, and some were still partly or wholly bound after ninety days. The authors read the pattern as evidence that the hydrolysis is enzymatic and not simply chemical, and note that the freed compounds go on to be converted into one another, changing the aroma profile further.

A study of curing beans found that glucovanillin moves from the inner part of the bean towards the outside during curing and is hydrolysed as it goes, with enzymatic hydrolysis taking place at the surface. Transcripts of a β-glucosidase gene from colonising microorganisms were detected, and the colonisers belonged mainly to the genus Bacillus. In culture, the isolates consumed glucovanillin as their only carbon source within 24 h, and vanillin was the main volatile produced. The authors conclude that Bacillus colonising the bean produce β-glucosidase that takes part in glucovanillin hydrolysis.

Pods killed in four different ways and then cured alike ended with different amounts of vanillin. Freezing gave the most, 1.89 per cent of dry tissue, followed by sonication at 1.56 per cent, hot water at 1.33 per cent and microwave at 1.14 per cent. Glucovanillin stood at about 1.3 per cent after freezing, sonication and hot water but at 3.49 per cent after the microwave treatment, so the microwave-killed pods had released the least of their precursor. The authors relate the result for freezing to mechanical rupture of cell walls and organelles.

Applying microwaves or ultrasound at the killing stage was reported to raise vanillin yield significantly and to reduce curing time to 20 days. Pods killed by ultrasound and given a further sonication at 38 °C yielded more than twice the vanillin of the control, but when that further sonication ran to 15 min, inactivation of β-glucosidase lowered the final yield. No significant correlation was found between β-glucosidase activity and vanillin across the treatments.

Cured vanilla beans stored at four temperatures diverged with time. Vanillin was highest in beans held at −20 °C for six months, with no perceptible change of colour. At 35 °C, products of lipid breakdown, enzymatic oxidation and the Maillard reaction, among them 2-acetylpyrrole and 5-hydroxymethylfurfural, were detected after 18 months, the colour had changed perceptibly, and the odour profile no longer met the standard the authors applied. Beans kept at −20 or 4 °C stayed within its colour requirements. Nonanal and 3,5-octadien-2-one were put forward as indicators of oxidation.