MtCROMlXlNG IN A STATIC MIXER AND AN EMPTY TUBE BY A CHEMICAL METHOD

Micromixing in a static mixer and an empty tube was characterized by the product distribution of fast consecutive competing reactions. As a test reaction, the precipitation of barium sulphate complexed by EDTA in alkaline medium under the influence of an acid was used. The acid was injected at a point of the reactor into an excess of basic complex. The precipitated BaSO4 is a measure for the degree of segregation which can be characterized by an index Xs representing the amount of macrofluid in the partially segregated fluid. Experiments were carried out in a tubular reactor and in a static mixer in the laminar flow regime (0.4 c Re ( 300) by varying the fluid viscosity (10e3 -Z u -Z 26-10e3 kg-m-l-s-t) and the flow rate. Furthermore, the characteristic reaction time tr was varied from about 1 to 75 s. Depending on the experimental conditions, the segregation index Xs varied from 0.1 to 0.3 for the mixer and from 0.2 to 0.75 for the empty tube. The micromixing time tm was deduced from Xs and tr. The mixing time tm was found to be about 10 times larger in an empty tube than in the static mixer. Introduction Tubular reactors for catalytic as well as non-catalytic reactions in homogeneous liquid phase are of special industrial interest. Polymerization reactions and homogeneously catalysed reactions involving soluble enzymes are examples among a broad range of applications. If media of high viscosity and/or slow reactions are involved, the tubular reactor should be operated in the laminar flow regime. As a consequence, its performance is impaired by the occurrence of unfavourable flow profiles and by convective fluctuations, due to temperature or density gradients. Both effects lead to a broad residence time distribution. Different means have been used to circumvent the above mentioned problems. Cascading and compartimentation of reactors with or without stirring may be used in addition to tower packings of diverse shapes. A comparison of these devices on the behaviour of tubular reactors with low fluid velocity can be found in Langensiepen. 1980. Since normal tower packings such as rings and saddles may lead to the formation of dead-zones in the reactor and give rise to high pressure drop for viscous media, regularly arranged rigid packings with high void volume are preferred. In spite of the fact that motionless mixers may have some of the above mentioned disadvantages, they generally improve tubular reactor performance as was proved by two experimental studies: the enzymatic hydrolysis of lactose in whey (Flaschel et al., 1983), and the thermal bulk polymerization of styrene in a pilot plant static mixer (Nguyen et at., 1983). These experiments cover a broad range of viscosities from u = 1 mPa=s in the first example up to u = 10 Pa*s in the second one, corresponding to Reynolds numbers between 500 to IO-5 respectively. In order to characterize the macroscopic behaviour of the tubular reactor with static mixers, the residence time distribution was determined and evaluated using the axial dispersion model. For the motionless mixer studied (type SMX of Sulzer, Switzerland) with an internal tube diameter of dt = 40 mm, Flaschel et al., 1985. determined a P&let number of Pe = (u*dt)/Dax = 3.2 in the range of IO-5 4 Re < 102. This corresponds to a Bodenstein number of 80 = u*L/Dax = 80 for a lube of 1 m length or to approximately 40 ideal mixing cells per meter using the tanks-in-series model. Metzdorf et al.,1986 confirmed the plug flow behaviour of static mixers at higher Reynolds numbers (50 c Re < 1000) and higher tube diameters up to dt = 300 mm. These results prove that effective macroscopic radial mixing can be achieved using motionless mixers, but they do not yield any information on the quality of microscopic mixing. Mixing on a molecular scale in the reactor becomes important for systems in which the micromixing time and the characteristic time constant of the chemical reaction are of the same order of magnitude. For complex reactions, e.g., co-polymerization, the quality of micromixing will strongly influence the product quality in terms of the mean molecular mass, the molecular mass distribution, the composition of copolymers and the sequence of the monomer units. Theoretical discussions were recently published, for example, by Ranz, 1986, and Fiekts and Onino. 1987. The aims of this contribution are to present a suitable test system for characterizing segregation in tubular reactors operating at low Reynolds numbers and to determine the influence of motionless mixing elements on the micromixing time. Measurement Principles Fast consecutive competing reactions are currently used for characterizing segregation in homogeneous systems: