Synthesis of organofluorine compounds using a falling film microreactor : process development and kinetic modelling.
MetadataShow full item record
South Africa is rich in valuable ore, including fluorspar (calcium fluoride), a principle feedstock used to synthesize hydrofluoric acid and a wide range of other fluorochemicals. As part of a wider initiative to promote local beneficiation of fluorspar, the primary purpose of this investigation was the synthesis of two valuable fluorochemicals, namely 1,1,2,3,3,3-hexafluoropropyl methyl ether (HME) and methyl-2,3,3,3-tetrafluoro-2-methoxypropionate (MTFMP). A secondary objective was to develop and identify suitable kinetic models as well as the associated kinetic parameters for the systems by means of least squares regression of experimental data. Two reactors were used in this study: the first, a gas-liquid continuously stirred semi-batch glass reactor was used for preliminary investigations (prior to the commencement of this work) to investigate the synthesis of HME. The second, a falling film microreactor (FFMR), a much more efficient medium for gas-liquid reactions, was used for the kinetic study. The FFMR was chosen as it has remarkable high rates of heat and mass transfer allowing for stringent control of reaction conditions as well as allowing for continuous process operation to be achieved. HME was produced by the reaction of methanol and hexafluoropropene in the presence of potassium hydroxide. MTFMP was similarly synthesized by the reaction of methanol and hexafluoropropene oxide in the presence of potassium and sodium hydroxide. The HME system was initially investigated in the gas-liquid glass reactor where the presence of HME as well as by-products of the reaction were successfully identified and quantified using gas chromatography-mass spectroscopy and gas chromatography (with the internal standard method), respectively. Results revealed that the reaction was rapid and highly exothermic and brought about negligible solid formation which deemed the system suitable for a FFMR. The yield of HME in the glass reactor varied between 19.21 and 53.32% with respect to moles of hexafluoropropene gas introduced into the system. The yields of the by-products were also quantified and the yield of alkenyl ether was found to vary between 0.07 and 1.03% while alkyl tetrafluoropropionate varied from 2.42 to 5.10%. These experiments were conducted at reactor temperatures between 12 and 28 °C, hexafluoropropene mole fractions in the feed between 0.41 and 0.83 and an inlet potassium methoxide concentration between 0.40 and 0.80 mol∙L-1. The novel synthesis of HME and MTFMP using a FFMR was then undertaken with the reactor operating in counter current mode. A 4 factor circumscribed Box-Wilson central composite design was used to design experiments, the four factors of interest being reaction temperature, liquid flowrate, either potassium or sodium hydroxide concentration, and finally, hexafluoropropene (for the HME system) and hexafluoropropene oxide (for the MTFMP system) mole fraction in the reactant gas. HME yields were found to be greater in the FFMR varying from 11.60 to 71.47% with the yields of the by-products varying from 0.11 to 6.99% for the alkenyl ether and 0.75 and 6.24% for the alkyl tetrafluoropropionate. Experiments were conducted at temperatures between 2 and 22 °C, hexafluoropropene mole fraction in the feed of 0.17 and 0.88, potassium methoxide concentration of 0.25 and 0.61 mol∙L-1 and a liquid flow rate of 0.50 and 5.50 mL∙min-1. The MTFMP system was handled similarly with its presence in the product being identified and quantified using gas chromatography-mass spectroscopy and gas chromatography (with the internal standard method), respectively. The yield of MTFMP was found to vary between 0.00 to 23.62% with respect to the moles of hexafluoropropene oxide introduced over the reaction period. The low yields were due to an inherently slower reaction which was not ideal for a FFMR given its low liquid residence time. Experiments were conducted at temperatures between 30 and 40 °C, hexafluoropropene oxide mole fraction in the feed of 0.16 and 0.84, potassium methoxide concentration of 0.15 and 0.65 mol∙L-1 and a liquid flow rate of 0.50 and 5.50 mL∙min-1. Results of the experiments showed that the MTFMP system was not suitable for the FFMR due to the low residence time of the reactant in reactor. A kinetic model was then developed for both the reactive systems from fundamental principles and executed in the MATLAB® environment (version R2012b, The MathWorks, Inc.). The kinetic model proposed for the HME system consisted of five reactions for which reference kinetic rate constants and activation energies were successfully identified. The resultant model did not predict HME concentrations adequately with an average absolute relative deviation percentage (AARD %) of 34.12 %. The reaction mechanism proposed for the MTFMP system was a two step reaction sequence which described the system satisfactorily well, the kinetic parameters required for this system was a reference kinetic rate constant, activation energy and Sechenov coefficient which were all successfully identified. The effect of reaction order on the pertinent reactions was also investigated and it was found that a second order description best fit the experimental data. The MTFMP model fit experimental data less agreeably with an AARD % of 170.00 % which was heavily weighted by data points with errors in excess of 500%, analysing the results exclusive of these points lead to an AARD % of 17.80 %.