In a 1L three-neck flask, set up with an addition funnel, was added 29.5 g (0.142 mol) isodillapiole dissolved in 55 ml dichloromethane, followed by 7.1 g (0.08 mol) sodium bicarbonate. To the addition funnel was added a solution of performic acid, which was prepared an hour before by mixing 30.6 g (0.225 mol) 25% hydrogen peroxide and 29 ml (35 g, 0.645 mol) 85% formic acid. This was added dropwise to the stirring suspension in the flask, which caused a lot of foaming due to the liberation of CO₂. The addition took around 90 minutes, and the two-phase mixture was allowed to stir for an additional 16 hours. At the 3 hour point, a TLC was done on a sample (solvent: methanol, detection: iodine vapours) which showed a large spot - presumably of both the glycol and glycol formyl ester - and only a very minor spot of isodillapiole, so it is safe to say that most of the olefin was consumed at this point.

The organic layer was separated and washed twice with water. The dichloromethane was distilled off and the resulting dark-red oil was taken up in 40 ml methanol. To this was added 60 ml water, followed by 60 ml 29.5% hydrochloric acid. This mixture was refluxed for 3 hours and after letting it cool to room temperature, the dark oily precipitate was separated, and the mother liquor extracted once with 60 ml dichloromethane. The extract and the oil were pooled, washed twice with water and once with an aqueous NaHCO₃ solution. The washings and the mother liquor were pooled and a small amount of oil was separated off and added to the organic solution¹. The dichloromethane was distilled off, then vacuum was applied and 17 g of greenish-yellow 2,3-dimethoxy-4,5-methylenedioxyphenylacetone came over at 195°C at the fridge pump (50% overall yield)².

Notes:

1. TLC (solvent: petroleum ether, detection: iodine vapours) showed that all glycol (formyl ester) was consumed and the post-reaction extract gave a single spot of the ketone, at a somewhat lower R_f value than the former. It must be noted that both the glycol (formyl ester) and the ketone contained an impurity which did not elute with this solvent system. These are the R_f values for those who are interested:
   

   silica gel plates, solvent methanol:
   dillapiole: 0.58
   isodillapiole: 0.57
   isodillapiole glycol (formyl ester): 0.52 (taken during performic oxidation after 3 hours of stirring)

   silica gel plates, solvent petroleum ether:
   isodillapiole glycol (formyl ester): 0.07-0.15 (a streak, due to mixture of glycol and ester?)
   2,3-dimethoxy-4,5-methylenedioxyphenylacetone: 0.05
   

2. The somewhat low yield was due to the use of an insufficient vacuum, as during the whole distillation smoke was seen entering the receiver. The column was warmed with an electric insulating mantle to prevent massive polymerisation caused by heating the oil too long. To obtain better results, the boiling point of the ketone must be kept under at least 185°C, as was observed before by Antibody2. A dark, viscous, sharp-smelling residue was left in the distilling flask, which solidified on cooling. The colour of the ketone went from greenish-yellow to a more golden tint after standing an hour at room temperature.