Local structure of network modifier to network former ions in soda‐lime alumino‐borosilicate glasses

Abstract

The structure of soda‐lime alumino‐borosilicate glass was studied using molecular dynamics simulations of samples of varying compositions containing ~20 000 atoms each. Pair distribution functions (PDFs) of cations to oxygen were used for comparison to available experimental data to evaluate consistency between simulations and experiment. Additional PDFs and coordination of the network forming cations (Al/B/Si) to network modifiers (Ca/Na) were examined, which is difficult to measure experimentally. The results are consistent with available experimental data regarding cation‐oxygen bond lengths and network former to oxygen coordination numbers. Si and Al are predominantly 4‐coordinated, with a small concentration of overcoordinated species similar to experimental data. B varied as 3‐coordinated, BO3, and 4‐coordinated, BO4, as a function of the amount of Ca2+ and Na+ present, the ratio of Al2O3 to B2O3, and the fictive temperature of the sample, similar to experimental data. The simulations provide new information regarding the locations on the network modifiers to the +3 cations, Al and B. For instance, one Al ion can have multiple Na within 4 Å, but also the Na can be within 4 Å of several +3 cations. Such results would indicate a greater complexity of local structure that goes beyond the stoichiometric one +1 modifier ion near one +3 network former or one +2 modifier near two +3 formers in tetrahedral sites.

Simulations of the surfaces of soda lime aluminoborosilicate glasses exposed to water

“…The system sizes werẽ 65 A 9~65 A 9~65 A with~20 000 atoms each. While the quench rate used here (and previously 15 ) is fast (~10 13 K/s), we also redid some of the previous compositions with a quench rate of~10 11 Al5-B15 60 5 15 10 10 19 998 Al10-B10 60 10 10 10 10 19 998 Al15-B5 60 15 5 10 10 19 998 Ca20 60 10 10 20 0 20 000 Ca10 70 10 10 10 0 19 998 Na20 60 10 10 0 20 19 998 Na10 70 10 10 0 10 19 998 with no significant change in results. While the quench rate used here (and previously 15 ) is fast (~10 13 K/s), we also redid some of the previous compositions with a quench rate of~10 11 Al5-B15 60 5 15 10 10 19 998 Al10-B10 60 10 10 10 10 19 998 Al15-B5 60 15 5 10 10 19 998 Ca20 60 10 10 20 0 20 000 Ca10 70 10 10 10 0 19 998 Na20 60 10 10 0 20 19 998 Na10 70 10 10 0 10 19 998 with no significant change in results.…”

“…The compositions and labeling protocol for the rest of the paper are given in Table 1. 15 The 3-body term centered on Al used in the previous paper 15 had parameters for k ij , c ij , and R ij equal to 0.024 fJ, 2.8 A, and 3.0 A, respectively. This lack of difference in these quench rates was also observed by Adelstein and Lordi 27 in their simulations of multicomponent glasses.…”

“…The NBO species labeled xOCaNa are formers (x) that have both a Ca and a Na within the first O to modifier near neighbor peak in the pair distribution function. 15 Reiterating the important result previously presented regarding the presence of multiple modifiers near any particular +3 cation 15 is worthwhile here: while the simple stoichiometric concept of one +1 modifier ion charge compensating one +3 cation, or a +2 modifier charge compensating two +3 cations is correct, it misses the actual atomistic distribution of multiple modifier species around multiple +3 cations, as shown in Figures 6 and 7 from Ha and Garofalini, 15 which also means that there can be several +3 cations around any modifier ion. 15 Reiterating the important result previously presented regarding the presence of multiple modifiers near any particular +3 cation 15 is worthwhile here: while the simple stoichiometric concept of one +1 modifier ion charge compensating one +3 cation, or a +2 modifier charge compensating two +3 cations is correct, it misses the actual atomistic distribution of multiple modifier species around multiple +3 cations, as shown in Figures 6 and 7 from Ha and Garofalini, 15 which also means that there can be several +3 cations around any modifier ion.…”

“…[5][6][7][8][9][10][11][12][13][14] In a recent paper, 15 the bulk structure of aluminoborosilicate glasses containing 10%-20% Ca and Na modifiers was explored using molecular dynamics simulations of systems containing~20 000 atoms. A brief summary of the recent paper 15 indicates a number of simulation results that were consistent with experimental data of the structure of bulk glasses: the bond lengths of each cation to oxygen was consistent with experiment; an increase in the Al 2 O 3 / B 2 O 3 results in a decrease in the BO 4 (often called N4 or 4 B) concentration at constant modifier content, consistent with Wu and Stebbins 14 ; an increase in the Ca concentration caused an increase in the 5 Al and a decrease in the 4 B, similar to experiment 11,14 ; an increase in the Na concentration increased the 4 B, similar to experiment 12,14 ; using the local coordination number, Na preferentially associates with Al > Si > B, similar to experiment 16 ; higher charge Ca enables a reduction in Al or B avoidance, similar to experiment 17,18 ; a decrease in the glass density leads to a decrease in the 5 Al, similar to experiments on fictive temperature on 5 Al concentration. The compositions were not designed to match any particular experiment, since there are many with varied compositions.…”

Molecular simulations of the structure and thermal transport of high alumina aluminosilicate molten core glass fiber

“…In recent simulations of multicomponent silicate glasses, a small concentration of 5‐coordinated Al is observed . Parameters are the same as previously published …”