Antimicrobial Fe2O3-CuO-P2O5 glasses

Abstract

Glasses with high antimicrobial efficacy were developed in the Fe₂O₃-CuO-P₂O₅ ternary system to mitigate fomite-mediated transmission of infectious diseases in high-risk settings such as hospitals, daycares, and nursing homes. Binary CuO-P₂O₅ glasses were not durable enough for use as high touch point articles, so Fe₂O₃ was added to the compositions to increase the chemical durability. The amount of Cu leachate decreased by at least 3 orders of magnitude when Fe₂O₃ was increased from 0 to 13.1 mol%. At the highest Fe₂O₃ contents and corresponding highest durability, the glass was no longer able to pass a test of antimicrobial efficacy with < 3 log kill compared to > 5 log kill for all other compositions. Ab-initio molecular dynamics simulations showed increasing bridging oxygen species at the expense of non-bridging oxygen species with the increase in Fe₂O₃ content, showing that the glasses exhibited increased chemical durability because they were more interconnected and structurally bound. Experimental results with glasses at fixed CuO and decreasing Fe₂O₃ confirmed that Fe₂O₃ content (not CuO) controlled the Cu release rate and, thus, the antimicrobial efficacy of the glasses. The significance of the oxidation state of the leached Cu was overwhelmed by the importance of the amount of Cu leachate.

Figure 1

Ternary diagram (0.25Fe₂O₃–CuO–P₂O₅) showing the glass forming region in grey and the six glass compositions of interest.

Figure 2

(a) The amount of leached Cu decreases on a log scale as a function of increasing Fe₂O₃ content, emphasizing the increased aqueous chemical durability with Fe₂O₃ content. Glass #s are labelled above each composition. Data plotted is renormalized to exclude SiO₂. (b) Leached Cu also decreases as a function of % bridging oxygens, showing that as the glass structure became more interconnected (more bridging oxygens), less Cu could be released from the surface. Note that bridging oxygens are defined as oxygens that link neighboring P- and Fe-polyhedra. These values were calculated from the AIMD simulation results. Glass #6 is not shown as the modeled composition was modified to contain significantly less Fe₂O₃ than the experimentally obtained glass.

Figure 3

(a) Glasses with higher Fe₂O₃ contents leached less Cu on each sampling day, indicating greater aqueous chemical durability. More Cu was leached at lower Fe₂O₃ contents even if there was less total CuO in the composition (e.g., light purple squares with 41.0 mol% CuO compared to dark purple diamonds with 41.6 mol% CuO). (b) Glass #4 (1.8 ± 0.3 log kill) showed low AM efficacy (< 3 log kill) despite having a comparable or greater amount of total CuO than glasses #5 and #6. Day 1 data is shown as the most comparable leach data to the conditions of the modified U.S. E.P.A. test. Glass #s are labelled above each composition. Data plotted is renormalized to exclude SiO₂.

Figure 4

(a) Fe₂O₃ content (mol%) v. Cu¹⁺/Total Cu, Fe²⁺/Total Fe, and total reducing power $\left(\left[{\text{Cu}}^{1+}\right]+\left[{\text{Fe}}^{2+}\right]\right)/\left(\left[\text{Cu}\right]+\left[\text{Fe}\right]\right)$ showing that the highest Fe₂O₃ content glass (#4) was the most reduced. Despite being the most reduced, glass #4 was the only composition that did not have high AM efficacy. Glass #s are labelled next to the three values for each composition. Horizontal dashed lines are used to distinguish the data for glasses #3, #5, and #6 as they have similar Fe₂O₃ contents. Data plotted is renormalized to exclude SiO₂. Cu¹⁺/Total Cu is only 0.01 greater than the total reducing power for glass #2 because the amount of Fe is just 13% of the total moles of redox species. (b) Leached Cu (ppm) after Day 1 on a log scale as a function of Cu¹⁺/Total Cu. The glass with the most Fe₂O₃ contained the most reduced Cu, which should have had the highest AM efficacy. However, high AM efficacy was not observed with this composition likely due to the low amount of total Cu released, as shown in (b).

Figure 5

Ab-initio molecular dynamics simulations based on analyzed compositions. Glass #1 was not simplified, and glasses #2–#6 were simplified to obtain reasonable computation times. Glass #4 was also run without simplification to assess the validity of the simplification assumption (shown as the second #4 column). The AIMD results showed a more connected glass network at higher Fe₂O₃ contents via an increasing fraction of bridging oxygen species (P–O–P, Fe–O–Fe, and P–O–Fe) with a corresponding decreasing fraction of non-bridging oxygen species (P–O–x and Fe–O–x) as a function of Fe₂O₃ content for Series A glasses with fixed P₂O₅ at 45 mol% and (55-X_Fe2O3) mol% CuO and for Series B glasses with fixed CuO at 42 mol% and (58-X_Fe2O3) mol% P₂O₅.