Prediction Of Parameters Of Anodized TiO₂/ZrO₂Nanotubes Using Artificial Neural Network

As important functional structure materials, Titanium dioxide (TiO2) nanotube andzirconium oxide (ZrO2) nanotubes have prospective application in plenty of fields dueto theirs small size, large surface area, etc. excellent properties. How to prepare TiO2、ZrO2nanotubes which morphology and properties are controllable have a key role fornanotube’s performance and application, therefore, for a long time, researchers havebeen very concerned about the preparation of TiO2and ZrO2nanotubes and carried outa large number of experiments exploring. Although experimentally explored can beprepared a number of different sizes of nanotubes, it is still difficult to achievecontrollable preparation of nanotubes. Hence, the development of simple, rapid andaccurate method to predict the morphology parameters of nanotubes and guiding thepreparation of nanotubes controllability are very important and urgent.In this paper, artificial neural networks was used to train the existing literature data,and extracted out the relationship between tube parameters of nanotubes (length,diameter and wall thickness) and anodic oxidation conditions, it can predict themorphology and properties of TiO2nanotubes and ZrO2nanotubes that prepared byanodization, which not only solved the problem of time-consuming andmoney-consuming in experiments explorating, but also could guide the preparation ofcontrollability nanotubes. The main work includes：1. Through analyzing the anodic oxidation conditions (electrolyte composition,temperature, oxidation voltage, and oxidation time) on the parameters of TiO2nantube,the lectrolyte composition, temperature, oxidation voltage, and oxidation time weretreated as the input data, and considered the length, thickness and diameter as outputdata, after several calculations, finding the most suitable number of hidden layer nodesfor TiO2nanotube’s length/thickness/diameter were all3, and eventually extract out theprediction equation for the parameters of nanotubes as follow：Length/um=-102.14*H1+764.93*H2+94.38*H3+554.80；Diameter/nm=-124.36*H1-35.50*H2+273.75*H3-42.86；Thickness/nm=-26.31*H1-23.00*H2-29.82*H3+33.42The tube length、diameter and wall thickness of the random test results further showthat: the above prediction formula can correctly predict the TiO2nanotubes’ length、diameter and wall thickness. 2. Through analyzing the anodic oxidation conditions’ influence (electrolytecomposition, temperature and oxidation voltage) on the parameters of ZrO2nantube, thelectrolyte composition, temperature, oxidation voltage, and oxidation time were treatedas the input data, and considered the length and diameter as output data, after severalcalculations, finding the most suitable number of hidden layer nodes for TiO2nanotube’s length and diameter were5/3(inorganic) and3/4(organic), and eventuallyextract out the prediction equation for the parameters of nanotubes as follow：Length/um=-243.55*H1-81.96*H2-1.76*H3-100.26*H4-155.70*H5+7.60；Diameter/nm=342.84*H1+719.93*H2-339.21*H3+157.73；（Inorganic system）Length/um=10.51*H1+14.07*H2+10.26*H3+15.22；Diameter/nm=57.63*H1+29.54*H2-30.04*H3-7.45*H4+38.52（Organic system）The tube length and diameter of the random test show that: the above predictionformula can correctly predict the ZrO2nanotubes’ length and diameter.