This thesis concerns the production of natural compounds from plant material for pharmaceutical and food applications. It describes the production (extraction and isolation) of cannabinoids, the active components present in cannabis. Many cannabinoids have medicinal properties but not all cannabinoids are available in the (large) quantities necessary to develop new medicines, because so far, for large scale production, there are no economically and technically viable methods to extract those cannabinoids present in low quantities in the plant. Moreover, the currently used production process for the most important cannabinoid, tetrahydrocannabinol (?9-THC), has many drawbacks, such as the large use of the organic solvents, which is not only a burden to the environment but also to the safety of the operators, the production costs as well as the treatment of the produced waste. In this thesis, an alternative process using supercritical carbon dioxide is presented for the production of cannabinoids, including ?9-THC, cannabinol (CBN), cannabigerol (CBG) and cannabidiol (CBD). One of the steps of ?9-THC production from cannabis plant material, is the decarboxylation reaction, transforming the ?9-THC-acid naturally present in the plant into the psychoactive ?9-THC. Experiments showed a pseudo first order reaction, with an activation barrier of 85 kJ.mol-1 and a pre-exponential factor of 3.7x108 s-1. Using molecular modeling, two options for an acid catalysed ?-keto acid type mechanism were identified. Each of these mechanisms might play a role, depending on the actual process conditions. Formic acid was shown to be a good model for a catalyst of such a reaction. A direct keto-enol mechanism catalyzed by formic acid seems to be the best explanation for the observed activation barrier and the pre-exponential factor of the decarboxylation of ?9-THC-acid. Evidence for this was found by performing an extraction experiment with Cannabis Flos. It revealed the presence of short chain carboxylic acids supporting this hypothesis. Then, in order to develop the supercritical fluid extraction process, the solubility of ?9-THC, CBN, CBG and CBD in supercritical carbon dioxide has been determined using an analytical method with a quasi-flow apparatus. First the solubility of ?9-THC has been determined at 315, 327, 334 and 345 K and in the pressure range from 13.2 to 25.1 MPa. The molar solubility for ?9-THC ranged from 0.20 to 2.95x10-4. Then, the solubility of CBN, CBG and CBD in supercritical carbon dioxide has been determined at 314, 327 and 334 K and in the pressure range from 11.3 to 20.6 MPa. The molar solubility of CBN, CBG and CBD ranged from 1.26 x 10-4 to 4.16 x 10-4, from 1.17 to 1.91 x 10-4 and from 0.88 to 2.69 x 10-4, respectively. These solubility data have been compared to each other. The solubility of the different cannabinoids in supercritical CO2 increases at 326 K in the following order: ?9-THC < CBG < CBD < CBN. The solubility data were correlated using the Peng-Robinson equation of state in combination with Van der Waals mixing rules. To continue, supercritical fluid extraction (SFE) using carbon dioxide was performed with Cannabis Sativa L. in a pilot scale set-up at 313 and 323 K in the pressure range from 18 to 23 MPa. The SFE yield of ?9-THC is at maximum 98 %, which is comparable to classical hexane extraction. CBN and CBG can be extracted in higher amounts with SFE than with hexane extraction. Waxes are co-extracted with the cannabinoids. They can be easily removed via a winterization step. The purity of the final extract after winterization was 85 % ?9-THC at the optimal experimental conditions found in these experiments. With a two-steps extraction, it is possible to selectively extract minor cannabinoids (i.e. CBN, CBD and CBG) in a first step at low pressure (~15 MPa), and ?9-THC in a second step at higher pressure (~20 MPa). The last step of the process is performed using Centrifugal Partition Chromatography. It uses a two-phase liquid system, instead of a solid stationary phase, as it is the case in High Pressure Liquid Chromatography (HPLC). Separation is realized by the partitioning of compounds between the two phases. With this technique, a successful separation of ?9- THC, CBN and CBG is presented using the two-phase system hexane / acetone / acetonitrile. A purity higher than 99% is achieved with ?9- THC. With CBN and CBG the best purity obtained is higher than 90%. To conclude, an economical and ecological evaluation of two production routes to obtain pure ?9-THC is presented: the current process using organic solvents is compared with the alternative process using supercritical carbon dioxide developed in this thesis. The alternative process is significantly cheaper than the current one, although the high price of the starting material cannabis dominates the ultimate cost price. From an ecological point of view, the alternative process is also more sustainable as it consumes less energy and generates less waste. Therefore, this alternative process is preferred from an economical and ecological point of view.