Sodium Iron Phosphate Glasses Research Articles

Evaluation of GlassNet for physics‐informed machine learning of glass stability and glass‐forming ability

AbstractGlassy materials form the basis of many modern applications, including nuclear waste immobilization, touch‐screen displays, and optical fibers, and also hold great potential for future medical and environmental applications. However, their structural complexity and large composition space make design and optimization challenging for certain applications. Of particular importance for glass processing and design is an estimate of a given composition's glass‐forming ability (GFA). However, there remain many open questions regarding the underlying physical mechanisms of glass formation, especially in oxide glasses. It is apparent that a proxy for GFA would be highly useful in glass processing and design, but identifying such a surrogate property has proven itself to be difficult. While glass stability (GS) parameters have historically been used as a GFA surrogate, recent research has demonstrated that most of these parameters are not accurate predictors of the GFA of oxide glasses. Here, we explore the application of an open‐source pre‐trained neural network model, GlassNet, that can predict the characteristic temperatures necessary to compute GS with reasonable performance and assess the feasibility of using these physics‐informed machine learning (PIML)‐predicted GS parameters to estimate GFA. In doing so, we track the uncertainties at each step of the computation—from the original ML prediction errors to the compounding of errors during GS estimation, and finally to the final estimation of GFA. While GlassNet exhibits reasonable accuracy on all individual properties, we observe a large compounding of error in the combination of these individual predictions for the PIML prediction of GS, finding that random forest models offer similar accuracy to GlassNet. We also break down the performance of GlassNet on different glass families and find that the error in GS prediction is correlated with the error in crystallization peak temperature prediction. Lastly, we utilize this finding to assess the relationship between top‐performing GS parameters and GFA for two ternary glass systems: sodium borosilicate and sodium iron phosphate glasses. We conclude that to obtain true ML predictive capability of GFA, significantly more data needs to be collected.

Insights on the structure and properties of sodium iron phosphate glasses from molecular dynamics simulations

Insights on the structure and properties of sodium iron phosphate glasses from molecular dynamics simulations

Atomic structure of sodium iron phosphate glasses

AbstractThe atomic structure of a series of sodium iron phosphate glasses is studied using different experimental techniques: X‐ray and neutron diffraction (ND), infrared spectroscopy, extended X‐ray absorption fine structure (EXAFS), and X‐ray absorption near‐edge structure (XANES). Detailed information about the atomic pair correlations is obtained. The high resolution of ND in real space resolves two P–O distances at 1.48 Ǻ and 1.59 Ǻ as expected. All the glasses are found to consist of a phosphate tetrahedral network with metaphosphate chains and pyrophosphate units, and every phosphate unit is found to have two or three nonbridging oxygen (NBO) links available to coordinate with Na and Fe cations. The Fe–O coordination number in these glasses is found to decrease from 5.7 to 4.8 with increasing the Fe content, whereas the Na coordination number of approximately 5 is detected for all the samples.

Properties and Crystallization Behavior of Sodium Iron Phosphate Glasses

Ternary Na2O–Fe2O3–P2O5 (NFP) glasses with varying Na2O/Fe2O3, Na2O/P2O5, and Fe2O3/P2O5 ratios were prepared. The properties and crystallization tendencies were systemically investigated. It is shown that both density and chemical stability of the glass increase with Fe2O3. In contrast the Na2O/P2O5 ratio has little effect on the glass properties for a fixed Fe2O3 content. The crystallization behavior of the glasses was analyzed by DTA and XRD. Unlike Li2O–Fe2O3–P2O5 glasses NFP glasses were found to be stable against crystallization. 15Na2O–27Fe2O3–58P2O5 glass was found to have the highest chemical stability among the studied NFP samples; the influence of TiO2, ZrO2 on crystallization in this composition was studied. It is found that addition of 3.4 mol% TiO2 or 2.2 mol% ZrO2 had little effect on the crystallization behavior of this glass. However, when the amounts of TiO2 or ZrO2 were increased to 8.4 or 5.5 mol% respectively the glass readily devitrified. Furthermore the addition of fluorine (introduced by replacing Na2CO3 with NaF in the glass batch) leads to amorphous glasses which could be crystallized to form NaFeP2O7 upon controlled thermal treatment. With increasing NaF additions the activation emergy for crystallization decreased from 428 to 381 kJ/mol.

Insignificant mixed‐alkali effect in Li2O–Na2O–Fe2O3–P2O5 glasses

AbstractConductivity spectra of mixed‐alkali ions in iron phosphate glasses have been studied in the frequency range from 400 Hz to 200 kHz, and over a temperature range from 333 K to 493 K. The ac conductivity as a function of temperature was divided into two domains; one where the absolute of ac conductivity is close to the dc conductivity, and one where the absolute magnitude of the ac conductivity is larger than the dc conductivity. The concentration of charge carriers was found to be almost independent of temperature, but was weakly dependent on composition. It seems that the conductivity was essentially determined by the hopping frequency. The activation energy for the hopping frequency is close to that of the dc conductivity and the glasses behave as strong electrolytes. The frequency exponent appears to be independent of temperature, but depends very weakly on composition. The glasses show weak pronounced mixed‐alkali effect in the mixed lithium and sodium iron phosphate glasses. The small minimum and maximum in dc conductivity and activation energy, respectively, are further evidence of the low mobility of the lithium and sodium ions in these iron phosphate glasses. We have also presented a scaling of the conductivity spectra, which implies that the relaxation dynamics of charge carriers in the present glasses is weakly dependent on temperature and composition. (© 2007 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)

Electrical relaxation in mixed lithium and sodium iron phosphate glasses

The electrical conductivity of mixed-alkali, lithium and sodium, iron phosphate glasses has been studied in the frequency range from 0.2 to 200 kHz and over a temperature range from 323 to 493 K. The AC conductivity as a function of temperature was divided into two domains, one where the absolute magnitude of the AC conductivity was close to the DC conductivity and another where it was larger than the DC conductivity. The DC conductivity for the mixed-alkali, lithium and sodium, iron phosphate glasses is independent of the Li/Na ratio and there is no evidence of any mixed-alkali effect. The lithium and sodium ions have such a low mobility in both single- and mixed-alkali iron phosphate glasses that they make no detectable contribution to the total conductivity that is of electronic origin. Mössbauer spectral analysis indicates the presence of both Fe2+ and Fe3+ ions.

Synthesis and degradation of sodium iron phosphate glasses and their in vitro cell response.

The degradation profiles of six sodium iron phosphate glass formulations have been investigated using simple dissolution trials in deionized water. The glasses were produced from the appropriate phosphate salts by melting at 1200 degrees C in 5% Au/95% Pt crucibles. Dissolution rates varied from 0.2 gcm(-2)h(-1) for the 1% Fe glass to essentially zero over the 6-week test period for the 15% Fe and 20% Fe glasses. The overall degradation rate was found to vary according to the approximate relation: rate = 1.3e(-0.79x) gcm(-2)h(-1), where x is the percentage iron content of the glass. Glasses with 10% or greater iron content were observed to maintain a constant density over the course of the tests and thus appeared to degrade from the surface and not the bulk. In vitro cell response tests were conducted on the glasses using macrophages and primary craniofacial osteoblasts. These tests were performed on the glasses with 10% or greater iron content because glasses with lower iron content degraded too quickly. Confocal microscopy revealed a rounded macrophage morphology and IL-1beta production was low, suggesting little macrophage activation. However, a significant level of peroxide production was observed. Osteoblasts were observed to attach to the glass surfaces and spread, exhibiting a similar cytosketetal organization to the cells on the Thermanox controls, with a high level of F-actin organization. On balance, the 15% Fe glass performed slightly better than the 20% Fe glass in these assays.

Electronic conductivity in Na 2O–FeO–P 2O 5 glasses

dc and ac conductivity were studied in 50P2O5–(50−x)FeO–xNa2O and 50P2O5–(50−x)FeO–xCaO (7<x<20; in mol%) glasses. It is shown that the replacement of CaO by Na2O caused rather small increase in dc and ac conductivity. The conduction in these glasses is predominantly electronic controlled by electron hopping between Fe(II) and Fe(III) sites. Internal friction has been used to study the relaxation phenomena caused by electronic and ionic migration in sodium–iron phosphate glasses. Two peaks attributed to these processes were observed on internal friction spectra in the temperature range of 125–500 K at a frequency of about 1 Hz.

A study of polarization mechanisms in sodium iron phosphate glasses by partial thermally stimulated depolarization current

Abstract The electrical polarization of sodium iron phosphate glasses, containing from 5.3 to 14.9 mol% Na 2 O and with a constant Fe/P ratio of 0.67, was investigated by thermally stimulated polarization current (TSPC), and by depolarization current (TSDC) techniques. The partial TSDC measurements showed a group of distributed partial peaks, P 1 , whose maximum ranged from 186 to 227 K. A group of higher temperature peaks, P 2 , was found in a narrower temperature range from 257 to 264 K, with an additional process, P 2 ′ , at about 240 K. The process P 1 was caused by a dipolar relaxation showing a compensation with the compensation point at T c =247 K. The physical meaning of the compensation point could not be related to analogous phenomena in organic polymers. The higher temperature peaks represent space charge relaxations greatly affected by the increase in Na 2 O content and influenced by the remaining polarization and conductivity. The most likely cause for the electrical relaxation was an interaction between the Na + ions and the P–O − bonds. Increasing the Na + concentration caused an increase in the TSDC peaks and a slight increase in dc conductivity.

Properties and structure of sodium-iron phosphate glasses

Abstract Selected properties of phosphate glasses, containing from 14 to 43 mol% Fe2O3 and up to 13 mol% Na2O, have been measured. With increasing Fe2O3 and Na2O content, the density and dilatometric softening temperature increased, whereas, the thermal expansion coefficient and dissolution rate in water or saline at 90°C decreased. Glasses containing more than 25 mol% Fe2O3 had an exceedingly good chemical durability. Their dissolution rate at 90°C in distilled water or in saline solution was up to 100 times lower than that of window glass. Mossbauer and X-ray photoelectron spectroscopy indicate that iron(II) and iron(III) were both present in the glasses and the chemical durability improved with increasing iron(III) concentration. The outstanding chemical durability of these glasses was attributed to the replacement of POP bonds by more chemically resistant POFe(II) and POFe(III) bonds.

Thermally stimulated polarization and dc conduction in iron phosphate glasses

Thermally stimulated polarization (TSPC) and depolarization current (TSDC) techniques were used to study electrical polarization and conduction mechanisms in iron phosphate and sodium-iron phosphate glasses. TSDC measurements from 120 to 350 K show two current peaks, P1, attributed to the polarization caused by intrinsic dipolar defects, and P2, due to space-charge relaxation. The electrical conductivity was examined on the basis that the activation energy for electronic conduction is lower than that for ionic conduction. The dc conductivity depends upon iron oxide content and distance between iron ions, which suggests electronic conduction. The difference in activation energy between TSDC peaks and dc conductivity is discussed. Infrared absorption spectra indicate that iron ions can act as a network former and/or modifier depending upon the Fe(II)/Fe(III) ratio in the glass.

Temperature dependence of electrical conduction of sodium-iron phosphate glasses

Abstract The electric conductivity was measured from 150 K to 300 K on sodium-iron phosphate glasses having the composition of X Fe 2 + Y O 3 . (45 − X )Na 2 O·55P 2 O 5 ( X = 0−45, Y = 0.09–0.98) in which ionic and/or electronic conduction may be possible. The conduction mechanism was examined on the basis that the activation energy of electronic conducting glasses is generally lower than that of ionic conducting glasses, and the relation between the logarithm of conductivity and the reciprocal of temperature deviates from linearity at high temperatures ( T > θ D /2), θ D being the Debye temperature, due to the contribution of the variable-range hopping conduction. The results obtained are as follows: 1. (1) when X is less than 15, the electric behavior of the glass is similar to that of sodium phosphate glasses, which show the ionic conduction; 2. (2) when X is more than 15, the electric behavior of the glass is similar to that of iron phosphate glasses, which show electronic conduction; 3. (3) when X is 15, the transition of electric behavior depends upon the amount of Fe 2+ ions in the glass.