Role of cerium oxide in bioactive glasses during catalytic dissociation of hydrogen peroxide

The addition of cerium oxide to bioactive glasses, important materials for bone tissue regeneration, has been shown to induce multifunctionality, combining a significant bioactivity with antioxidant properties. We provide a real time investigation of the evolution of the electronic properties of highly diluted cerium ions in a liquid environment containing hydrogen peroxide - the most abundant reactive oxygen species in living cells. This challenging task is undertaken by means of high-energy resolution fluorescence detected by X-ray absorption near-edge spectroscopy at the Ce L3 edge. We investigate samples with variable compositions and different morphologies. We relate the observed spectroscopic changes not only to variations in the concentration of the two Ce oxidation states in the samples, but also to changes in the local atomic environment of Ce ions, providing a clear picture of the role of cerium ions in the dissociation of hydrogen peroxide. The obtained results contribute to the understanding of the mechanisms that come into play in the process and provide a basis for the optimization of the functionalities of this class of materials.


Specific surface area
The specific surface area (SSA) of the glasses were evaluated by adsorption of N 2 at a temperature of 77 K using a Micromeritics ASAP 2020 porosimeter.
The samples were outgassed in vacuum at room temperature for 12 hours before the measurements.
The BET model 1 was used to determine the surface area.
The results of the SSA measurements for the samples are summarized in table ST1

Determination of point of zero charge
The PZC (point of zero charge) value was determined for all the investigated samples using two methods: a) Simplified mass potentiometic titration method 2,3 Two identical solutions (blank and sample) were prepared with 3.0 mL of 0.1 M KNO 3 and 6.0 mL of deionized water, and their pH values were measured with a Conductronic 120 pH meter.1.0 mL of 0.01 M KOH were added to the blank solution and the pH was measured again.50 mg of the EC precipitate was then added to the sample solution, followed by 1.0 mL of 0.01 M KOH.Both the blank and the sample were then titrated with 0.01 M HNO 3 and the results were plotted.The PZC of each sample was estimated at the point where both titration curves crossed.
b) Salt addition method 4 This method consists in a simple titration that requires a smaller amount of solid sample than other methods.Here, 0.200g of each EC precipitate was added to 40.0mL of 0.1 M NaNO 3 in ten 50-mL plastic beakers.The pH was adjusted using a ThermoElectron Orion 4 Star pH meter to 2, 3, 4, 5, 6, 7, 8, 9, 10 and 11 (± 0.1 pH units) with 0.1 M HNO 3 and 0.1 M NaOH as needed in each beaker.These were then shaken for 24 h in a revolving water bath to reach equilibrium (Gyratory water bath shaker G76).After this time each resulting pH was measured and the initial pH (pH 0 ) vs. the difference between the initial and final pH values (pH) was plotted.The PZC was taken as the point where pH= 0.

Goodness-of-fit parameters
The fitting program used for data fitting, Fitxk 6 , uses the weighted sum of squared residuals as a  2 function of merit for the fit.This is defined as: where are the data points, are the fitting points and are the standard deviations.
where is the average of the data.̿

Figure S1 :
Figure S1: The degradation of H 2 O 2 was determined by soaking the glasses in a 0.1 M water solution (glass mass/solution volume = 5 mg/mL) in a stirrer and by determining the residual H 2 O 2 concentration by titration with KMnO 4 .The figure reports the residual H 2 O 2 concentration after 1, 2, and 4 h and 1, 4, and 7 days by the H (blue), K (red), MBG (orange) and MBG 4+ glasses (pink).The decrease is much faster for mesoporous glasses (MBG and MBG 4+ ) than for glasses obtained by melting (H and K).The H and K samples have a comparable dissociation rate, although the K sample contains a lower molar concentration of cerium oxide (3.6 %) compared to the H sample (5.6 %).The data on H glass are reproduced from reference 5 .

Figure S2 :
Figure S2: Ce L 3 -edge HERFD-XANES spectra for the two reference samples for Ce 4+ (a) and Ce 3+ (b) (solid lines) after subtraction of the edge-jump modelled as an arctan function.The spectra were acquired on a CeO 2 sample and on a cerium nitrate hexahydrate (Ce(NO 3 ) 3 •6H 2 O) sample, respectively, in the form of powders pelleted with cellulose, with a cerium oxide concentration comparable to the one of glass samples.The individual Gaussian fitting components (solid blue) and the overall fits (dashed lines) are also shown.The A 1 peak in Ce 3+ and the B 1 peak in Ce 4+ are very close in energy, therefore only the A 1 component was used in the glass spectra fitting.The amplitude of the B 1 component was instead fixed to ratio B 1 /(B 2 +C 1 +C 2 ) in the Ce 4+ reference, the amplitude of B 2 , C 1 and C 2 being fitting parameters.

Figure S3 :
Figure S3: Ce L 3 pre-edge features of Ce 3+ (red) and Ce 4+ (dark green) reference samples and of the H (green), K (purple), MBG (orange), and MBG 4+ (blue) glasses measured in pure water before the reaction.The pre-edge structures of the H and K samples exhibit a similar shape, with a dominant peak energetically close to the Ce 4+ pre-edge peak (5.722 keV) and a further minor feature, close to the dominant peak of the Ce 3+ reference spectrum (5.719 keV).In the MBG and MBG 4+ samples the Ce 3+ related peak has a progressively lower intensity compared to the H and K samples.The differences in the shape of the pre-edge features in the different samples are in qualitative agreement with the ones observed in the XANES region, thus supporting the simplified approach of considering the XANES of the glass samples merely as a superposition of Ce 3+ and Ce 4+ related components.

Figure S4 :
Figure S4: Evolution of the Ce L 3 HERDF-XANES spectra in the pre-edge region of H (a), K (b), MBG (c), and MBG 4+ (d) glasses during the reaction with a 0.1 M H 2 O 2 solution.The modifications induced in the pre-edge region by the reaction, although less evident than the ones observed in the edge region, are in agreement with a progressive oxidation of the samples, with the feature at 5.722 keV showing a mild increase and the one at 5.719 keV showing a mild decrease of intensity as the reaction proceeds.

Figure S5 :
Figure S5: Evolution of the Ce L 3 -edge HERFD-XANES spectra of the MBG glass during the reaction with a 10 M H 2 O 2 solution (top) and difference between the spectra after 140 min and before the reaction (bottom) for: a) the full XANES energy region; b) the pre-edge region.The spectral modifications are expectedly more significant than in the case of a 0.1 M solution.The differencespectrum in panel a shows a negative peak at 5.727 KeV, ascribed to shift of the edge jump to higher photon energies as the reaction proceeds, consistent with a mild oxidation.The two positive peaks correspond to increases in intensity at photon energies close to B 2 and C 1 .The modifications of the preedge peak are limited to a very small shift to lower photon energies, consistent with a mild oxidation.

Table ST1 :
Specific Surface Area of the investigated glass samples

Table ST2 :
 2 values of the fittings of the different samples at different reaction times

Table ST3 :
R 2 values of the fittings of the different samples at different reaction times