Design of CeyCoxTi (1-x) O3+δ Perovskite for Photocatalysis: A Statistical Study

Abstract

The influence of metal composition and preparation pH on the properties and photoactivity of sol-gel synthesised Ce

Photocatalytic Degradation of p-Cresol by Zinc Oxide under UV Irradiation

“…where, the estimated value of k cresol (mg L −1 min −1 ppm −1 ) and K cresol (ppm −1 ) were 0.0403 and 0.33 with R-squared value 0.98 respectively. Similar results were reported for photodegradation kinetics of aqueous sodium oxalate, sodium dodecyl sulfate, 4-nitrophenol solution using TiO 2 photocatalyst [ 33 – 36 ] and our previous work [ 25 ].…”

A new weighted optimal combination of ANNs for catalyst design and reactor operation: Methane steam reforming studies

“…here, q ¼ 1, 2, …, NE h2 denotes the number of neurons in the second hidden layer (h2) and b h3 r represents the bias applied to the hidden neuron q. 6. Similarly, for the third hidden layer the input and output for the r th neuron is given by…”

“…3 As a result, ANN modeling and analysis has emerged as a versatile tool for data treatment and utilization in the nonlinear processes typically encountered in rational catalyst design and optimization as well as reactor engineering. 6 In many ANN modeling studies, the optimal network architecture was often selected based on the candidate with the smallest value of one of a class of error indices, such as the residual sum of squares errors (SSE), the mean square error (MSE), the root-mean square error (RMSE), the mean absolute error (MAE), and the mean absolute percentage error (MAPE). More recently, Chesterfield and Adesina 5 also obtained the functional relationship between titania photocatalyst preparation variables, its physiochemical properties and photoactivity using a 16neuron single hidden layer feed forward ANN based on over 700 data cases from different laboratories over a 20-year period.…”

“…Availability of the quantitative relationship between photocatalyst preparation variables such as calcination tem-perature, dopant ionic radius, oxidation number, etc., and the photoactivity for organic pollutant destruction has led to the design of new visible-light activated codoped titania for photocatalysis. 6 In many ANN modeling studies, the optimal network architecture was often selected based on the candidate with the smallest value of one of a class of error indices, such as the residual sum of squares errors (SSE), the mean square error (MSE), the root-mean square error (RMSE), the mean absolute error (MAE), and the mean absolute percentage error (MAPE). However, since different error indices measure different extents and types of mismatch between the ANN model and the data, it is rare that a single ANN will have the smallest value for all five types of error indices from among a set of ANNs originally culled for consideration.…”