high-yield production of amorphous carbon via ball milling of graphite and prediction of its crystallite size through ann

One of the main methods for the synthesis of amorphous and nanostructured carbon is the mechanical milling of graphite. however, calculation and anticipation of the amorphous phase during the mechanical milling of graphite still is a major challenge due to a lot of important parameters. the main aim of this study is to mass-produce amorphous carbon and predict the crystallite size of graphite. for this purpose, ball-milling of graphite powder was carried out at different times of milling. then, the destruction of crystal structure and changes in phases were studied by xrd, tem, afm, sem, and zeta seizer. the results of themaud analysis showed that 91% and 93% of the unmilled graphite were converted to amorphous carbon at 250 and 330 hours of ball-milling, respectively. in order to predict the crystallite size of carbon during the high energy ball-milling, the effective variables in the ball-milling process along with the initial crystallite size of carbon were determined as the input of the artificial neural network (ann). moreover, the final crystallite size of carbon was considered as the output of the network. the designed network with a root mean square error (rmse) of 4% was able to predict the crystallite size of carbon during the process. finally, by comparing the experimental results and the designed model, it was shown that the predicted results were very close to the experimental outcomes. accordingly, the presented model can be used for predicting the crystallite size of carbon during the mechanical milling of graphite.