Chemical Engineering Journal, Vol.238, 17-26, 2014
A new method for studying activity and reaction kinetics of photocatalytic water oxidation systems using a bubbling reactor
A novel method is proposed for studying kinetics and overall activity of water oxidation (WO) catalysts using a bubbling reactor, where oxygen concentration is measured simultaneously in the liquid and in the gaseous phase. Total oxygen evolution is obtained from direct integration. The actual rate of oxygen formation as a function of time, R-o2 (t) not accessible to direct measurement with batch reactors, is calculated from raw data through a simple but comprehensive mathematical model, taking into account mass transfer phenomena occurring in the system. Data concerning the activity of a nanostructured Co3O4 catalyst dispersed on a mesoporous silica (MSU-H), in the presence of tris(2,2'-bipyridyl)Ruthenium [Ru(bpy)(3)](2-) as sensitizer and Na2S2O8 as sacrificial reactant, are used to illustrate data processing. Behaviour of the system is complicated by the occurrence, besides WO reaction, of the degradation of the sensitizer. Increase of sweeping gas flow increases R-o2 (t), by decreasing diffusional limitations to the reactions in the system: conditions for minimizing those were established. Data reported show that the assumption generally made of equilibrium between gaseous and liquid phase through Henry's law is incorrect, the more so the smaller the apparent mass transfer coefficient, k(L)a. An additional reason for removing oxygen from the liquid phase through bubbling is the occurrence of a parasitic reaction of molecular oxygen with the sensitizer. The method seems to yield reliable values of both IcLa and the set of R-o2 (t) values: the former scales with the flow of sweeping gas, as expected; R-o2 (t) curves are in qualitative agreement with accepted reaction mechanisms. Results concerning R-o2 (t) lend support to our previous kinetic studies (M. Armandi et al., ACS Catal, 2013, 3, 1272) where the reaction rate was assumed as constant for the first similar to 15 min. Availability of R-o2 (t) data not too biased by diffusional limitations opens the way to realistic studies of the kinetic features of WO heterogeneous reactions, in the present case as well as in many others. (C) 2013 Elsevier B.V. All rights reserved.