Nathan Barrowa, Michael Packardb, Shuchi Vaishnavc, Paul Binghamc, Martin Wildingc, Alex Hannond, Ian Slagleb, and Steve Fellerb
a. Johnson Matthey Technology Centre, Blount’s Court, Sonning Common, Reading, RG4 9NH, UK
b. Physics Department, Coe College, Cedar Rapids, IA 52402, USA
c. Materials and Engineering Research Institute, Sheffield Hallam University, Sheffield, City Campus, Howard Street, Sheffield, S1 1WB, UK
d. ISIS Facility, Rutherford Appleton Lab, Chilton, Didcot, Oxon, OX11 0QX, UK
This work was supported by the NSF under grant number DMR-1746230 and the Johnson-Matthey Company.
A report of various structural properties is given over a wide compositional range of sodium silicate glasses. This investigation covers two sets of samples, one made with naturally abundant carbon and one enriched with 99.9% 13C. All NMR samples were doped with iron oxide (Fe2O3) to decrease NMR relaxation times. Both sets of samples were synthesised under comparable situations.
Amorphous samples of nominal Na2O-SiO2 compositions of 20, 25, 30, 35, 40, 45, 50 mol% Na2O were produced by conventional melt quenching. Glass samples of nominal compositions of 55, 60, 65, and 70 mol% were made using rapid quenching roller quenchers, extending the known glass-forming range for this composition.
From 29Si MAS NMR a Q-Unit distribution was found that follows the expected lever-rule trend, except at high sodium composition.
This deviation from the trend is due to carbonate retention, which essentially under-modifies the samples. 13C, 23Na, and 29Si MAS NMR was performed and the following observations were made: structural evolution of the Q-Units was confirmed over the entire range of compositions studied. 13C NMR confirmed the retention of carbonate at high alkali concentrations in form of carbonate, as well as sodium carbonate at 70 mol%. This observation was confirmed in the 23Na NMR in which two phases of sodium were observed for 70 mol%. At lower sodium concentrations the structure is almost identical, implying similar structures.
Carbon retention of this glass system was quantified by LECO analysis. At nominal compositions 20, 25, 30, and 35 there was no more than 0.01% carbon detection. In higher compositions, the carbon detection increases with composition. In the compositional range between 40 mol% and 55 mol%, all values for carbon detection were between 0.036% and 0.118%. Towards the end of the glass forming range, carbon detection was considerably higher. We see this in the highest three compositions by soda content: at 60 mol% the carbon detection was 0.27%, at 65 mol% the carbon detection was 1.47%, and at 70 mol% the detection was 3.05%.
Using the carbon detection and NMR, a net sodium modification was calculated that represented the actual modification of the glass, compensating for carbon retention. For nominal compositions lower than 60 mol%, there was little deviance between the expected and measured compositions, with the percentage of NMR derived sodium oxide to expected soda at near 99%. At the nominal compositions of 60, 65, and 70 molar percent Na2O there was much deviation from the expected amount of sodium oxide modifying the glasses due to the carbon dioxide retention. From 60 mol% to 70 mol% the percentage of net modifying Na2O to batch drops from 96.7% to 66.5%, indicative of strong CO2 retention.
The density of these glasses follows the changes in composition as follows: there is an upward trend in density up to the nominal composition of 55 molar percent soda. Above this composition the density decreases due to the presence of carbonate retention.