Experiments for final optimization of conditions and the results

As mentioned above, some of the primary optimal conditions determined in the first and second stages cannot yet be confirmed as really optimal, because they are the boundary values tested. In order to optimize these conditions further, some supplementary experiments were carried out in wider ranges to examine the influences of the pH, the reaction temperature, and the concentration of TiCl4 solution on the mean size of the product, for which the uniform design technique was employed. The experimental conditions and the results are listed in Table 15.8, where the values for the average sizes of particles are calculated from the X-ray spectrums. 

Level pH Temperature °C Concentration of TiCl4, kmol-m Average size, nm Yield of Ti, % 

A non-linear regression analysis of the data listed in the table is made by Li [211]. By comparison between the regression coefficients the following are concluded (1) The pH has the most significant effect on the mean size, the influence of the TiCI4 concentration takes second place, while the temperature has little effect and (2) Similarly, the pH and the TiCl4 concentration exhibit the most significant influences on the yield of titanium. 

From the point of view of the influences on both the mean size and the yield of titanium, it seems that the pH could be lower however, excessively low pH may aggravate corrosion of the equipment, while an excessively high concentration of TiCl4 would result in difficulties of operation. Combining these considerations, the optimal values finally determined for these three parameters are as follows  

---