Experiments were conducted in two systems. The first, an all-glass system with a static reaction vessel of volume 167 cm3 and surface to volume ratio of 1.00 cm-1, has already been described (Stark and Waddington, 1995). The second system, capable of operating at much higher pressures, consisted of either a 1 or 2 litre stainless steel (316) autoclave operated as a continuously stirred tank reactors (CSTR) with a residence time of 30 - 100 s. Mixing was ensured by stirrers, magnetically driven at 1500 rpm and was monitored by two thermocouples located at the top and bottom of the reactor volume. A third thermocouple monitored the reactor wall temperature. The gases were premixed and preheated to ca 370K before entry into the reaction vessel, and the exit gas was kept at ca 390K until passing through the back pressure regulator.
The analysis technique for the static system, principally by gas chromatography, has been described earlier (Stark and Waddington, 1995). That for the flow system was similar, except that oxygen and nitrogen were determined using a 30m 5A molecular sieve PLOT column with oxygen consumption verified by paramagnetic oxygen analyzers (Servomex 1100H).
For the flow system, an ice trap collected the easily condensable products. The cold trap liquids were treated with standard sodium hydroxide solution using phenolphthalein as indicator to determine the amount of organic acids present. The infrared spectrum of the dried salts showed that the principal acid was formic acid (ca 95% (v/v)), with the remainder mainly acetic acid. The condensate in the cold trap also contained quantities of propane-1,2-diol, 4-methyl-1,3 dioxolane, propane-1,2-diol mono- and di-formates and propane-1,2-diol polymers. These species are not thought to be formed in the gas phase, but by heterolytic cleavage of the epoxide bond in the highly acidic condensed phase. They are recorded as propene oxide derivatives and the sum of these species plus the epoxide is taken as a measure of the epoxide formed in the gas phase.
The product yields reported here are expressed as selectivities (the fraction of the propene reacting going to a particular product). For the propene:acetaldehyde co-oxidation experiments only the increase in concentration of the acetaldehyde was allowed for in the calculations. The quoted epoxide selectivities were reproducible to better than +-5% points.
The concentration of acetaldehyde added in these experiments was chosen so that its concentration at the end of the reaction was approximately equal to that at the beginning. This occurs when the quantity consumed is equal to that produced by the reaction of propene and oxygen, ie. the concentration of acetaldehyde that would occur in a flow system that had the acetaldehyde separated out of the exit gas and recycled to the input of the reactor (figure 1). The reaction numbering used in this paper is consistent with the published reaction scheme which has subsequently been used to simulate the experiments reported here (Stark and Waddington, 1995).