ABSTRACT
The photo-Fenton process was applied to degrade non-ionic surfactants with different numbers of ethoxy groups, seven (E7), ten (E10) and twenty-three (E23). The effects of H2O2 concentration, Fe(II) concentration and number of ethoxy groups on the mineralization of surfactants were investigated. The response surface methodology (RSM) was applied to determine optimal concentrations of Fenton's reagents for each surfactant. The efficiency of the photo-Fenton process reached 95% for all surfactants studied at 45 min in optimal conditions determined in this work. The analysis of results showed that the efficiency depends upon the number of ethoxy groups in the surfactant. The increase in ethoxy groups favoured the mineralization of surfactants. The analysis of variance (ANOVA) was applied, and according to the F-test the models for the mineralization of surfactants were considered significant and predictable. The photo-Fenton process has proven to be feasible for the degradation of ethoxylated surfactants in aqueous solution.
Subject(s)
Ethyl Ethers/chemistry , Fatty Alcohols/chemistry , Photolysis , Surface-Active Agents/chemistry , Hydrogen Peroxide/chemistry , Iron/chemistryABSTRACT
An artificial neural network (ANN) was implemented for modeling phenol mineralization in aqueous solution using the photo-Fenton process. The experiments were conducted in a photochemical multi-lamp reactor equipped with twelve fluorescent black light lamps (40 W each) irradiating UV light. A three-layer neural network was optimized in order to model the behavior of the process. The concentrations of ferrous ions and hydrogen peroxide, and the reaction time were introduced as inputs of the network and the efficiency of phenol mineralization was expressed in terms of dissolved organic carbon (DOC) as an output. Both concentrations of Fe(2+) and H2O2 were shown to be significant parameters on the phenol mineralization process. The ANN model provided the best result through the application of six neurons in the hidden layer, resulting in a high determination coefficient. The ANN model was shown to be efficient in the simulation of phenol mineralization through the photo-Fenton process using a multi-lamp reactor.
Subject(s)
Bioreactors , Light , Models, Theoretical , Neural Networks, Computer , Phenol/chemistryABSTRACT
The application of the photo-Fenton process for the treatment of wastewaters contaminated with diesel oil was investigated. This particular process has been widely studied for the photochemical degradation of highly toxic organic pollutants. Experiments were performed according to a factorial experimental design at two levels and two variables: H(2)O(2) concentration (5-200 mM) and Fe(2+) concentration (0.01-1 mM). Experimental results demonstrated that the photo-Fenton process is technically feasible for the treatment of wastewaters containing diesel oil constituents, with total mineralization. A combination of factorial experimental design and gradient descent techniques was employed to optimize the amount of the Fenton reagents, resulting in Fe(2+) (0.1 mM) and H(2)O(2) (50 mM). These optimized levels did not exceed the limit for disposal of ferrous ions (0.27 mM) proposed at the local environmental legislation.
