High-level nuclear waste (HLW) from defense programs is immobilized in borosilicate glass, and in high-alumina Hanford glasses the main threat to chemical durability is nepheline (NaAlSiO4) crystallization during canister cooling, which depletes the residual glass of network formers. McClane and co-workers (2020) showed that durability after crystallization is governed by the residual glass composition and its network connectivity, and identified four linked gaps that must be closed before the conservative compositional constraints on Hanford HLW glasses can be relaxed: a mechanistic prediction of surface versus bulk nepheline nucleation under realistic canister thermal gradients; quantification of how spatially heterogeneous crystallization alters local residual-glass composition, diffusion, connectivity, and element-specific leaching; an a priori method to choose compositional adjustments that steer residual glasses away from deleterious phase fields across Hanford's diverse waste compositions; and validated laboratory-to-canister scaling rules that correct specific-surface-area biases. Later work (2021-2026) built occurrence predictors, a residual-glass durability screen, and identified surface-area-to-volume ratio as the main scaling confounder, but all four gaps remain open, and 2026 studies still describe the prediction-boundary and scaling problems as ongoing.

Background

Nearly all high-level radioactive waste (HLW) from U.S. defense nuclear programs — including the tank waste at the Hanford site — is immobilized by vitrification into borosilicate glass, in which waste oxides are dissolved in a SiO2-B2O3-Na2O-Al2O3 matrix. The disposal safety case rests on the chemical durability of the glass, i.e., its resistance to leaching of radionuclides and toxic elements in water, routinely benchmarked by standardized leach tests. Because Hanford wastes are alumina-rich, the most important constraint on how much waste can be loaded into a given glass (the waste loading, which sets the number of canisters and the mission cost) is crystallization of nepheline (NaAlSiO4) during the slow cooling of a large canister, the thermal history conventionally specified as canister centerline cooling (CCC). Nepheline removes Na2O, Al2O3, and SiO2 from the melt, and the leftover (residual) glass is then depleted in network formers; this residual glass, not the starting composition, controls the measured leach response. Glass network connectivity is described by Q units (the number of bridging oxygens around a network-forming tetrahedron, Q4 = fully connected, Q0 = isolated tetrahedron) and by the non-bridging oxygen (NBO) content; lower connectivity generally means faster dissolution. McClane, Amoroso, Fox, Hsieh, Kesterson, and Kruger (Int. J. Appl. Glass Sci. 11, 649-659, 2020, DOI 10.1111/ijag.15418) quantified how compositional shifts of the residual glass after nepheline crystallization — tracked through calculated Q distributions — produce distinct deviations in leaching behavior, and stated the practical goal: broadening the acceptable glass compositional region so Hanford HLW can be vitrified more efficiently. Their predecessor study (McClane et al., J. Non-Cryst. Solids, 2019, DOI 10.1016/j.jnoncrysol.2018.11.018) had characterized nepheline crystallization kinetics in simulated HLW glasses. In their conclusions, the authors identify four unresolved issues that block immediate generalization and deployment of relaxed compositional rules. First, there is no mechanistic, predictive understanding of the nucleation pathway — when nepheline nucleates at the sample surface versus homogeneously in the bulk — under the real thermal gradients of a cooling canister. Second, crystallization is spatially heterogeneous, so the residual glass develops local compositional gradients whose effects on local diffusion, network connectivity (NBO and Q distribution), and element-specific leaching are not quantified. Third, there is no a priori method to predict which compositional adjustments steer the residual-glass composition away from deleterious phase fields — regions of the Na2O-Al2O3-B2O3-SiO2 phase space, such as the NaAlB2O5-type region, whose crystallization products degrade durability — across the broad diversity of Hanford HLW compositions. Fourth, laboratory durability and crystallization measurements use small specimens whose specific surface area (surface-area-to-volume ratio, SA/V) is much larger than that of a full canister, and robust scaling rules or experimental protocols that correct this SA/V bias so laboratory data can forecast full-canister crystallization and durability have yet to be developed and validated. Subsequent work built on but did not close these gaps: Vienna et al. (Int. J. Appl. Glass Sci., 2021, DOI 10.1111/ijag.16074) and Lu, Sargin, and Vienna (J. Am. Ceram. Soc., 2021, DOI 10.1111/jace.17983) developed submixture models (empirical predictors based on ratios of oxide submixtures rather than single oxides) and machine-learning predictors of nepheline occurrence reaching about 90% accuracy; McClane et al. (J. Am. Ceram. Soc., 2021, DOI 10.1111/jace.17768) derived the Structural Integrity of Residual (SIR) glass model, which screens post-crystallization durability from the calculated NBO content of the residual glass and identifies durable glasses containing 20-35 wt% Al2O3; Krishnamurthy, Michaelis, and Kroeker (J. Phys. Chem. C, 2021, DOI 10.1021/acs.jpcc.1c01262) and Deshkar et al. (J. Non-Cryst. Solids, 2022, DOI 10.1016/j.jnoncrysol.2022.121694) advanced the structure-property linkage; and the 2025-2026 Hanford-focused studies of Cornwall, Scrimshire, Bell, Smith, Kruger, Bingham and co-workers (Phys. Chem. Glasses, 2025, DOI 10.13036/17533562.66.4.am36; Int. J. Appl. Glass Sci., 2026, DOI 10.1111/ijag.70034) traced predictor failures near the nepheline boundary to SA/V differences between small samples and canisters, while Sargin (Int. J. Appl. Glass Sci., 2026, DOI 10.1111/ijag.70055) showed that onset and amount of nepheline are governed by distinct mechanisms and that existing binary models are insufficient.

Problem Statement

Establish a mechanistic and predictive framework for nepheline-controlled durability of Hanford high-level waste borosilicate glasses — sufficient to relax the current conservative compositional constraints with confidence — by resolving four linked questions posed by McClane et al. (2020): (1) predict, from composition and thermal history, the nepheline nucleation pathway under realistic canister cooling gradients, i.e., when crystallization initiates at the surface versus homogeneously in the bulk; (2) quantify how spatially heterogeneous crystallization generates local residual-glass compositional gradients, and how those gradients alter local diffusion, network connectivity (non-bridging oxygen content and Q-unit distribution), and element-specific leaching of the residual glass; (3) predict a priori which compositional adjustments steer the residual-glass composition away from deleterious phase fields (such as the NaAlB2O5-type region) across the broad diversity of Hanford HLW compositions; and (4) develop and validate scaling rules or experimental protocols that correct laboratory specific-surface-area (SA/V) biases so small-specimen crystallization and durability data reliably forecast full-canister behavior under canister centerline cooling. These four questions are jointly required rather than independently sufficient: a result for any proper subset of them is partial progress toward, but does not by itself complete, the task, and the framework counts as established only once all four results are obtained and jointly provide a defensible, validated basis for accepting higher-waste-loading, high-alumina glass compositions.

References

  1. McClane, Amoroso, Fox, Hsieh, Kesterson, Kruger (2020). Nepheline crystallization and the residual glass composition: Understanding waste glass durability. International Journal of Applied Glass Science 11(4), 649-659. DOI: 10.1111/ijag.15418. https://doi.org/10.1111/ijag.15418
  2. McClane, Amoroso, Hsieh, Fox, Kruger (2021). Derivation of the Structural Integrity of Residual (SIR) glass model for the enhancement of waste loading. Journal of the American Ceramic Society 104, 3235-3246. DOI: 10.1111/jace.17768. https://doi.org/10.1111/jace.17768
  3. Vienna, Kroll, Hrma, Lang, Crum (2021). Submixture model to predict nepheline precipitation in waste glasses. International Journal of Applied Glass Science. DOI: 10.1111/ijag.16074. https://doi.org/10.1111/ijag.16074
  4. Lu, Sargin, Vienna (2021). Predicting nepheline precipitation in waste glasses using ternary submixture model and machine learning. Journal of the American Ceramic Society 104, 5636-5647. DOI: 10.1111/jace.17983. https://doi.org/10.1111/jace.17983
  5. Krishnamurthy, Michaelis, Kroeker (2021). Network Formation in Borosilicate Glasses with Aluminum or Gallium: Implications for Nepheline Crystallization. Journal of Physical Chemistry C 125, 8815-8824. DOI: 10.1021/acs.jpcc.1c01262. https://doi.org/10.1021/acs.jpcc.1c01262
  6. Deshkar, Parruzot, Youngman, Gulbiten, Vienna, Goel (2022). Compositional dependence of crystallization and chemical durability in alkali aluminoborosilicate glasses. Journal of Non-Crystalline Solids 590, 121694. DOI: 10.1016/j.jnoncrysol.2022.121694. https://doi.org/10.1016/j.jnoncrysol.2022.121694
  7. Cornwall, Scrimshire, Bell, Smith, Kruger, Bingham (2025). On the thermal and kinetic methods of nepheline precipitation within Hanford high-level nuclear waste canisters. Physics and Chemistry of Glasses: European Journal of Glass Science and Technology Part B 66(4). DOI: 10.13036/17533562.66.4.am36. https://doi.org/10.13036/17533562.66.4.am36
  8. Cornwall, Bell, Scrimshire, Smith, Filleul, Sargin, Kruger, Bingham (2026). On the Conservatisms of Nepheline Predictors for Canister Centerline Cooled (CCC) High-Level Waste (HLW) Glasses. International Journal of Applied Glass Science. DOI: 10.1111/ijag.70034. https://doi.org/10.1111/ijag.70034
  9. Sargin (2026). Domain-Knowledge-Guided Rule Extraction for Governing Nepheline Amount in High-Level Waste Glasses. International Journal of Applied Glass Science. DOI: 10.1111/ijag.70055. https://doi.org/10.1111/ijag.70055
  10. Gan, Gilbo, Pegg, Joseph (2026). Nepheline constraint for Hanford HLW glass. U.S. DOE / OSTI report. DOI: 10.2172/3021669. https://doi.org/10.2172/3021669
  11. McClane, Amoroso, Fox, Kruger (2019). Nepheline crystallization behavior in simulated high-level waste glasses. Journal of Non-Crystalline Solids 505, 215-224. DOI: 10.1016/j.jnoncrysol.2018.11.018. https://doi.org/10.1016/j.jnoncrysol.2018.11.018