Concerns regarding the migration of potentially hazardous elements from glass containers into their contents, particularly lead from lead crystal glassware, have driven stricter regulations. This study aims to comprehensively investigate the long-term durability of common commercial glass types under accelerated aging conditions. Four glasses were selected to represent a range of compositions and applications: lead crystal (Glass A), known for its lead content and potential leaching; barium crystal (Glass B), a lead-free alternative; soda-lime glass (Glass C), widely used in food and cosmetic packaging; and borosilicate glass (Glass D), employed in heat-resistant cookware. Understanding the alteration mechanisms and durability of these glasses is crucial for ensuring food contact safety and informing future regulations. The study will assess how composition, network structure, and alteration conditions influence leaching behavior and the formation of alteration layers. This knowledge will contribute to the development of more durable and safer glass materials for various applications, promoting consumer health and sustainable practices.

Previous research highlighted the significant lead migration from lead crystal glassware (Graziano et al., 1991; Hight, 1996; Guadagnino et al., 2000). European Directive 1969 defined 'crystal' glass as containing at least 24 wt% lead oxide, raising health concerns. Barium crystal glass emerged as a potential safer alternative (European Directive). Studies have explored the alteration mechanisms of soda-lime and borosilicate glasses, focusing on factors like pH and temperature (Perera & Doremus, 1991; Sinton & LaCourse, 2001). Research using solid-state NMR has investigated glass structure and its influence on durability (Angeli et al., 2006; Angeli et al., 2016; Tricot, 2016). However, a direct comparison of these four distinct glass types under identical, long-term conditions has been lacking, making this study a significant contribution to the understanding of glass durability.

Four commercial glasses—lead crystal (A), barium glass (B), soda-lime glass (C), and borosilicate glass (D)—were analyzed. Their compositions were determined using ICP-AES after complete dissolution (Table 1). Glass powders (63–125 µm) and polished slabs (20 × 20 × 2 mm³) were subjected to an accelerated aging test in 4% (v/v) acetic acid (pH 2.4) at 70 °C for 1096 days (3 years). The leaching solutions were sampled regularly and analyzed by ICP-AES to monitor alteration kinetics. The glass-surface-area-to-solution-volume ratio (SA/V) was calculated using both geometric approximation (Eq. 1) and BET surface area measurements. The normalized mass loss (NL) for each element was calculated (Eq. 2), and converted to equivalent thicknesses (ETh) of altered glass (Eq. 3). Alteration rates were derived (Eq. 4), and diffusion coefficients for elements exhibiting interdiffusion behavior were calculated (Eq. 5). Data consistency was ensured using correction factors (Eqs. 6 and 7). After 231 days, glass slabs underwent ToF-SIMS and spectroscopic ellipsometry (SE) analyses. Solid-state NMR spectroscopy (¹¹B, ²³Na, ²⁷Al, and ²⁹Si MAS NMR) characterized the pristine and altered glass structures. SEM-EDX analyzed polished cross-sections of altered glass grains.

The study revealed significant differences in the alteration mechanisms and rates of the four glasses. Glass A (lead crystal) showed a rapid initial leaching of alkalis (Na, K), followed by a decrease in the leaching rate due to the formation of a passivating layer. Lead was initially released rapidly but then became largely retained in this surface layer, likely through the formation of Si-O-Pb bonds. SEM-EDX and ToF-SIMS confirmed the depletion of alkalis and the retention of lead in the altered layer. ²⁹Si MAS NMR indicated silicate network repolymerization, contributing to lead retention. Glass B (barium glass) exhibited a much lower alteration rate than Glass A, with minimal differences in constituent leaching. The added alkaline earth elements strengthened the glass network. Glass C (soda-lime glass) showed preferential Na release, consistent with ion exchange/interdiffusion. A slight decrease in alteration rates was observed after 300 days, suggesting alteration layer formation. SEM-EDX revealed a surface layer depleted in sodium. Glass D (borosilicate glass) displayed an almost congruent dissolution, with similar alteration rates for all elements. No significant alteration layer was observed, indicating that hydrolysis dominated the alteration process. ToF-SIMS depth profiles revealed variations in alteration layer thickness among glasses (Table 2). Glass A displayed the thickest altered layer (7 µm), while Glass D showed a minimal altered layer (<50nm). The rate of silicon hydrolysis remained remarkably similar across all glasses despite variations in composition and network structure, suggesting the hydrolysis rate is primarily dictated by experimental conditions (pH, temperature, and SA/V). The leaching of sodium was strongly linked to the polymerization degree of the glassy network, with the least polymerized glass (A) exhibiting the highest Na release rate.

The findings highlight the strong influence of glass composition and structure on alteration behavior. The significantly different alteration rates observed between the glasses underscore the importance of compositional control in designing durable and safe glass materials. The consistent silicon hydrolysis rate suggests that this process is primarily governed by the experimental conditions, while the rate of alkali leaching is influenced by network polymerization and the formation of alteration layers. The repolymerization observed in Glass A demonstrates a dynamic interplay between alteration mechanisms, potentially leading to self-healing properties. The data from ToF-SIMS and SE provide valuable insights into the nanoscale structure and composition of the altered layers, complementing the information from solution analyses. The observed differences in alteration between glasses B and C suggest that barium contributes more to durability than magnesium. The results support the use of highly polymerized glasses for applications requiring high durability, like borosilicate glass in cookware.

This study offers a comprehensive understanding of the long-term durability and alteration mechanisms of four common glass compositions. The differences observed in alteration behaviour are largely determined by differences in glass composition and network structure. The finding of a similar silicon hydrolysis rate despite large differences in glass composition suggests that this process is determined primarily by experimental parameters. This research has implications for the glass industry and regulatory bodies. It underscores the need to consider long-term durability in glass material design and strengthens the case for using highly polymerized glasses in applications requiring high chemical resistance. Further research could focus on investigating the effect of other environmental factors on glass alteration and exploring new glass compositions with enhanced durability and safety.

The study used accelerated aging conditions to shorten the experimental timeframe. While this allowed for a thorough investigation of long-term effects, it might not perfectly replicate real-world conditions. The use of acetic acid to simulate food and beverage content might not capture the full complexity of real-world interactions. Some minor lab contamination (calcium, potassium) was observed and dealt with through data corrections, potentially introducing slight uncertainty in the results.