Theoretical Studies On The Micro-structure Of Molten Silicates And Its Relation With The Macro-properties

It is the terminal intention of this thesis to link the microstructure of silicate melts to their macro-properties. So the first thing is to correctly acquire the entire knowledge about the microstructure. In these years only two ways are feasible to directly measure the microstructure of molten silicates at 2000K. The high temperature Raman spectroscopy (HTRS) is one of them. Generally, experiments only can give some discrete information. The calculation with the model based on the discrete information, after it has been verified by experiments, is the integrated knowledge. Furthermore, the illustration of some details and the quantitative analysis of the HTRS of molten silicates are still a hard nut to crack so far. Therefore, the construction of the micro-structural model of silicate melt is the key to realize the terminal purpose of this thesis. The author recognizes that the Raman spectrum is essentially a distribution of the molecular vibrational behaviors. According to this recognition, in order to make a correct micro-structural model for melt it is necessary to perform at first a statistic analysis of the configurations of the melt. This is carried out by means of MD (molecular dynamics) simulation and its attached function module. The MD simulation results indicate that for any two random Si-O tetrahedra belonging to the same kind, their homologous bond lengths and bond angles are different with each other because of their different microenvironments. The statistical results are starting point of carrying the theoretical calculation, including the eigen vibrational analysis, the intensity computation and the envelope integration, of Raman spectra for melt. All these are included in the SiOT (Si-O tetrahedra) model, and the results calculated with this model are compared with the measurements of HTRS. The last chapter is the CEMS (Cluster-Equilibrium of Molten Silicate of Metallurgical Slag) model. Based on either calculated or measured micro-structural information, this model can work out the mixing free energy of molten silicates and the relative thermodynamic properties.Some of the significant points in this thesis should be elucidated in detail as follows:1 The MD simulation of microstructure of molten silicatesThe microstructure of three binary silicates CaO-SiO2, Al2O3-SiO2, CaO-Al2O3 and a ternary CaO-Al2O3-SiO2 was studied by means of classical MD simulation. In this simulation the empirical BMH pair potential was adopted, every sample got 630-1000 atoms, and the simulated ensemble was NPT or NVT. The rescaling or the Nose-Hoover thermostat methods were used to control temperature, and the pressure was controlled mainly by the Parrinello-Rahman method. The time step was settled as 0.001ps or 0.002 ps. The partial RDFs and CNs were achieved at first in the structural analysis. Through these two curves the average pair distances and coordinate numbers were extracted. The results were compared with the experiments to examine whether the simulation was successful. Following that, the resulted equilibrium configurations were decomposed to obtain the distributions of bond length, bond angle, three kinds of oxygen, and tetrahedral units. The agreement of the simulation results with the associate experiment results evidenced the reliability of the simulation and the feasibility of the application of BMH pair potential.The following information has been gained from MD simulation: i). In the studied system, Si4+ is network former; its coordinate number is 4. Ca2+ is network modifier, its coordinate number is 6 but not very stable. Al3" is an amphoteric ion; generally it plays the role of network former with coordinate number 4; if the silicate containsmuch Si4+ and little Ca2+, it may play the role of network modifier with coordinate number of 5 or 6, and usually A1V1 scarcely emerges, ii). The bond length d(Si-Ob) does not be effected by the change of composition, however, the d(Si-Onb) is distinctly effected. It becomes longer following the increase of CaO content, iii). The influence...