Optimization of cellulose phosphate synthesis from oil palm lignocellulosics using wavelet neural networks

Cellulose phosphate was synthesized from microcrystalline cellulose derived from oil palm lignocellulosics via the H3PO4/P2O5/Et3PO4/hexanol method. The influence of process variables (viz. temperature, reaction time, and the H3PO4/Et3PO4 ratio) on the properties of the resulting cellulose phosphate was investigated using a wavelet neural network model with the goals of ascertaining which factors were critical and of determining optimized reaction parameters for this synthesis. The experimental results corroborated the good fit of the wavelet neural network model. The prediction errors were quite small (less than 7%), and the regression values (R2 greater than 0.99) were also satisfactory.