The synthesis of R-heptenone is a two-step reaction with R-ketone as key intermediate. The first step is a Grignard reaction where R-butenol is formed from R-ketone. In the second step R-butenol and methyl acetoacetat form cis and trans R-heptenone through a [3, 3] sigmatropic rearrangement reaction. R-heptenone is then purified by distillation. However, gas chromatography shows two peaks with almost the same retention time as cis and trans R-heptenone, which is not eliminated in the distillation process. The main objective of the diploma work was to identify and reduce the amount of these impurities to the same level as achieved by another producer. The peak before cis-R-heptenone turned out to be two different impurities with identical retention times, the alcohol of cis-R-heptenone and a ditrans isomer of R-heptenone. It was shown that the peak slightly ahead of trans-R-heptenone as well originated from two substances, the alcohol of trans-R-heptenone and one still unidentified impurity. The formation of alcohols is catalyst by aluminium isopropoxide (Aliso P), the catalytic agent in the [3, 3] sigmatropic rearrangement reaction. However, the alcohols only forms if the amount of methyl acetoacetate is low compared to R-heptenone (<0.4 eqv). The ditrans isomer of R-heptenone arises from the ditrans isomer of R-butenol which itself seems to originate from the corresponding R-ketone isomer. A quality slightly better then R-heptenone from “other company” was obtained with the use of R-butenol with a lower level of the ditrans isomer and by making sure that the methyl acetoacetate level not was below 0.4 equivalent at any time. In practice, this means that the total amount of byproducts was decreased from 0.39 area percent (s = 0.0341, n = 46) to 0.14 area percent. In contrast to Aliso P Aluminum-tri-tert-butoxide (purity 97 %) lacks a-protons and showed a significantly weaker tendency to form the alcohols of R-heptenone. The formation of alcohols that did occur is most likely due to impurities (isomers) in Aluminum-tri-tert-butoxide. The yield with aluminum-tri-tert-butoxide turned out to be very good, 96 %, which is better than any other experiment (average = 89%, n = 34).