The instant application claims the benefit of US Provisional Patent Application Serial Number 62/291 ,727, filed June 29, 2018 and entitled “INCREASED MOLYBDENUM AND SULFUR SOLUBILITY IN ALUMINOBOROSILICATE GLASSES WITH ADDED PHOSPHORUS”, the entire contents of which are incorporated herein by reference.

BACKGROUND OF THE INVENTION

Spent fuel rods from the nuclear power plants are reprocessed in order to extract fissile materials which enter a new cycle of nuclear fission process. The remaining waste from the reprocessing step, rich in radioactive fission products, must be carefully handled to prevent harmful environmental exposure. The best way to insulate these radioactive wastes from the environment is to immobilize them in matrices characterized by long-term chemical durability and other desirable physical properties. Vitrifying calcined nuclear waste is the current practice among the nuclear power harnessing countries. Glass has been globally accepted for this task due to the salient features it possesses. Specifically, borosilicate glass is the de-facto

composition for nuclear waste immobilization [1][2] However, the major drawback of borosilicate glass is its low affinity for molybdenum, rare-earth elements and platinoid elements (Ru, Rh, Pd) which are highly concentrated in calcined nuclear wastes [3]. The high cation field strengths inherent to these fission products (for example, Mo ⁶⁺ has a field strength of 1.89-1.93 A ² [4]) induce a strong ordering of oxygen anions about the cation, ultimately leading to nucleation and crystallization which phase separates from the glass.

Molybdenum is one such fission product which phase separates by

sequestering alkali (A ⁺ ) and alkaline earth metal ions (B ²⁺ ) from the waste stream in the form of crystalline molybdates (A2M0O4 and BMo0 ₄ ) [4] This phase-separation phenomenon alters the structure and decreases the strength of these glasses, compromising their long-term performance. Mo ⁶⁺ is a key concern because of its ability to sequester isotopes with very high radioactivity (e.g., ¹³⁷ Cs and ⁹⁰ Sr) and long half-lives (e.g., ¹³⁵ Cs) to form complex molybdate assemblages sometimes called yellow phase [5] These radioactive crystalline phases are highly soluble in water and hence could leach into the environment upon contact of repository-stored wasteforms with water. M0O3 is one of the limiting factors which controls nuclear waste loading capacities in borosilicate glasses owing to its low solubility. Homogeneous glasses without any phase separation can be achieved when the M0O3 loading limit is below about 1 mol% (3 wt%) [6]

Like molybdenum, sulfate anions in radioactive waste have low solubility in borosilicate glasses (0.6 mol% or ~ <1 wt%) [7-9]. The sources of sulfate in a typical nuclear waste stream are ferrous sulfamate (used in the reduction of Pu ⁴⁺ to Pu ³⁺ )

[10], waste liquors arising from the purex process, and spent-ion-exchange resins

[1 1]. Though the oxidation states of sulfur range from -2 to +6, in the oxidizing conditions of borosilicate glasses, they are predominantly found in the +6 oxidation state [10] with a coordination number of four (SO4 ^2' ) [12]. Above the solubility limit, a sulfate-rich salt layer is observed on top of the melt pool, which precipitates during melt cooling [8] This phase-separated layer is called yellow phase [9] or gall [10] and is a well-known problem to be avoided in borosilicate glasses used in the

immobilization of high-level liquid wastes. The main component of the gall layer is Na2S0 ₄ [10], which is capable of sequestering radioactive isotopes of Cs, Sr and Tc [7,8,10,13] Phase separation and precipitation of the gall layer on top of the melt also has a detrimental effect on the crucible and furnace, as it is highly corrosive and, being a good conductor, makes the melting process less efficient [7] The gall layer on the melt pool inhibits the release of gas bubbles, leading to swelling of the vitreous phase upon cooling [10] Sulfate is introduced as isolated SO4 ^2' tetrahedral units into borosilicate glasses [12] Sodium sulfate is the main sulfate-bearing phase in a phase- separated borosilicate glass, and sulfate units are concentrated into voids in the borosilicate network [14] The chemical environment of sulfate in a borosilicate glass is similar to that of molybdate. The molybdate units are present as isolated tetrahedra concentrated in the depolymerized regions rich in alkali and alkaline-earth ions within the borosilicate glass [15]. This tendency to concentrate in the depolymerized regions rich in alkali ions leads to the eventual sequestration of alkali ions by the phase-separating molybdate units when Mo is loaded above its solubility limit [16]. The separated molybdate phases are distributed in borosilicate glasses as nano- and micrometer spherical phases and are known to follow nucleation-and-growth mechanism during phase separation [17]. Owing to the fact that both Mo ⁶⁺ and S ⁶⁺ are high-field-strength cations and are distributed within the borosilicate glass network as isolated tetrahedral units, it is reasonable to assume that they both follow similar alkali ion sequestration and phase-separation mechanisms.

Increases in M0O3 and SO3 solubilities will lead to more efficient and

economical immobilization processes, and reduce the volume of waste glass generated. Glass-melting temperatures are also important because at high

temperatures, sulfate immobilized in the glass decomposes, releasing SO2 gas which might be enriched with radioactive isotopes [18]. The development of new borosilicate glass compositions which can melt at lower temperatures, yielding homogeneous glasses and incorporating higher amounts of M0O3 and SO3 than currently possible, would have a significant impact on nuclear waste disposal worldwide, as discussed herein.

Phosphate glasses are known to incorporate higher amounts of M0O3 [19] and SO3 [20] than silicates. Although phosphate glasses have higher SO3 loading capacities, binary Fe203-P20s glass, which has been proposed as a potential glass composition for nuclear waste immobilization, has very low SO3 solubility (<0.1 mol%) [21 ,22]. The low sulfate solubility in highly durable iron-phosphate glasses makes other phosphate systems appear less desirable, even though they can incorporate high amounts of SO3 owing to their poor chemical durability compared to iron phosphate and borosilicate glasses. Phosphate glasses proposed for Mo

incorporation also suffer from poor chemical durability [23]. Since phosphate glasses are capable of dissolving high amounts of M0O3 and SO3, doping borosilicate glasses with P2O5 could potentially improve the solubility of these species in borosilicate glasses. However, very little has been published on this topic. Molybdenum phase separation was studied in calcium-sodium silicate glasses which contained up to 5 wt% of phosphorus and 12 wt% of molybdenum [24] Partial crystallization was witnessed in the glass and it was observed that molybdenum favored phase separation, whereas phosphorus contributed to both phase separation and crystallization. In another study, molybdenum immobilization was evaluated in silicate-phosphate glasses [23]. Two series of silicate-phosphate glasses were studied wherein one was silica-rich (silico-phosphate) and the other was phosphorus- rich (phospho-silicate). The glasses were melted at 1450°C and rapidly quenched in water. Molybdenum solubility between 4.4 and 5.7 mol% was measured in silico- phosphate glass which had 6 mol% of P2O5. Although higher Mo solubility was achieved, the glass consisted ~45 mol% of alkali and alkaline earth oxides [A2O +

AO], making it inappropriate for nuclear waste immobilization. Also, the glass did not contain any B2O3, which makes it less popular as the most widely used glasses for nuclear waste immobilization are borosilicates. A thirteen component glass system possessing 2 mol% P2O5 was studied by Raman spectroscopy and rheological measurements [25,26] Glasses were initially plate-quenched, then reheated and poured onto copper rollers, slow cooled at 60°C/h and finally rapidly quenched by air- blowing [26] Irrespective of cooling rates, phases with similar compositions separated from the glass and were determined to follow a nucleation-and-growth mechanism. The immiscibility range for this glass system was found to be between 1 150°C and 1200°C during which phases rich in Ca, Mo, P, Nd and O separated from a residual matrix rich in Si, Al, Zn, Zr and O. Unfortunately, the fate of Na ⁺ ions was not discussed, which might have provided insight into the structural roles of the highly radio-active ¹³⁵ Cs and ¹³⁷ Cs isotopes. These phase-separated droplets coalesced to form larger phases. The composition of phases separated from the glasses studied by Raman spectroscopy and rheology were similar. The rheology study was focused on relating the melt rheological behavior and phase separation process, and also to determine phase separation temperature.

The phase diagram of a six-component borosilicate glass with molybdenum and phosphorus was studied in detail by Pinet et. al. [27] The glasses were melted at 1300°C and rapidly cooled. Four different series were formulated, ranging from low to high silica content (32 to 44 wt%). Glasses with the highest aluminum and lowest phosphorus concentrations were homogeneous while low aluminum and high phosphorus yielded stratified glasses consisting of two superimposed opaque glass layers, one rich in phosphorus and the other rich in silica. Homogeneous glasses contained lower amounts of Si as compared to glass compositions adopted for nuclear waste immobilization, which could prove detrimental to their chemical durability. Since only the high-alumina glasses were homogeneous, the required melting temperatures were very high due to its high refractory content, making the process less economical.

Significantly, the aforementioned reports treated molybdenum and phosphorus only as waste components, with no apparent intention to improve Mo solubility by adding phosphorus. To our knowledge, no papers have been published wherein the researchers have tried to improve Mo solubility alone or to improve loading capacities of molybdenum and sulfate simultaneously as a function of phosphorus content in borosilicate glasses. However, it is known that the high field strength of Mo ⁶⁺ has an ordering effect on neighboring oxygens and results in sequestration and crystallization [16,28]. Generalizing this concept to a multicomponent glass with high-field- strength cations such as P ⁵⁺ , S ⁶⁺ and Mo ⁶⁺ will create competition for oxygens, resulting in more oxygen sharing and a greater degree of cation integration into the glass network. In other words, the high field strength of network-forming phosphorus should prevent the clustering and nucleation of Mo0 ₄ ² and S0 ₄ ² species into crystalline phases, thereby improving their incorporation into a homogeneous and durable bulk glass.

Boron also plays an important role in optimizing nuclear waste glass properties. It is a very good network former and forms homogeneous glasses when added to silicate and phosphate glasses due to favourable B-O-Si [29] and B-O-P [30] bonds. Boron is also known to bond well to intermediate network formers like aluminum through B-O-AI bonds [31], and has been proposed to connect to molybdenum through bridging oxygens, forming B-O-Mo bonds [19]. Similar to aluminum [27], boron can also be expected to interconnect silicon, phosphorus and molybdenum within the same glass network. Hence, by designing a boron-rich borosilicate glass, the problem of high-refractory glasses can be reduced and lower melting

temperatures can be used, solving the problem of radioisotope evaporation and making the melting process more economical.

As discussed herein, three sets of samples based on the SON68 glass composition [32] are prepared to compare the performance of different compositions with respect to the P-free aluminoborosilicate base glass. Initially, the solubility of molybdenum in borosilicate glass is established by varying the phosphorus content added to the borosilicate glass. Upon reaching the desired solubility of molybdenum, sulfur is added to test its solubility in the presence of molybdenum, mimicking the actual nuclear waste containment scenario. In series A, phosphorus (P2O5) is substituted for S1O2 at 0, 2.5, 5.0 and 7.5 mol%. Series B features a constant phosphorus content (10 mol%), with the Mo content increased from 3 to 5 mol%.

Series C contains 3.0 mol% of both M0O3 and SO3, while the amount of phosphorus is increased from 0 to 5 mol% at the expense of S1O2.

SUMMARY OF THE INVENTION

According to a first aspect of the invention, there is provided a composition for the preparation of borosilicate glass comprising:

30-60 mol% S1O2;

10-30 mol% B2O3;

2-30 mol% P2O5;

0-15 mol% AI2O3;

10-35 mol% Alkali oxide or alkaline earth oxide.

In some embodiments, the composition comprises: 45-55 mol% S1O2;

15-20 mol% B2O3;

2-10 mol% P2O5;

0-5 mol% AI2O3; and

15-20 mol% Alkali oxide or alkaline earth oxide.

According to another aspect of the invention, there is provided a modified borosilicate glass, for example an aluminoborosilicate glass, characterized in that up to 10 mol% of the silica has been replaced with P2O5.

According to another aspect of the invention, there is provided an aluminoborosilicate glass characterized in that up to 10 mol% of the silica has been replaced with P2O5.

According to a further aspect of the invention, there is provided a method of loading a borosilicate glass with nuclear waste comprising:

30-60 mol% S1O2;

10-30 mol% B2O3;

2-30 mol% P2O5;

0-15 mol% AI2O3; and

10-35 mol% Alkali oxide or alkaline earth oxide;

adding to the composition a powder comprising fission products in oxide form, thereby forming a mixture;

melting the mixture to a temperature of approximately 1100°C for sufficient time to melt the mixture; and

pouring the mixture into a suitable container. BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 : ²³ Na MAS NMR spectra of series A glasses.

Figure 2: ²³ Na MAS NMR spectrum of glass A-1 with fit, showing the presence of crystalline phases. The fit parameters are listed in table 2. Figure 3: X-ray diffractrogram of glass sample A-1 showing sharp peaks from the crystalline phase and the halo from the glassy phase. The reflections are assigned to sodium molybdate form in hydrated and dehydrated forms.

Figure 4: ¹¹ B MAS NMR spectra of series A.

Figure 5: ¹¹ B MAS NMR spectra of series A in absolute intensity mode.

Figure 6: Deconvolution of ¹¹ B MAS NMR spectra of glass samples from series A. Dotted red line represents the envelope from the fitting components. Quasar and gaussian/lorentzian models were used to fit ^[3] B and ^[4] B units respectively. The fit parameters are listed in Table 3.

Figure 7: Fraction of total boron in tetrahedral coordination, N ₄ , in series A glasses plotted as a function of phosphorus content. Dotted line is a guide for the eye.

Figure 8: Relative integral of different boron species present in glasses from series-A plotted as a function of phosphorus content in the glass.

Figure 9: ³¹ P MAS NMR spectra of glasses from series A (Example 1), glass B- 1 from series B.

Figure 10: ³¹ P MAS NMR spectra in absolute intensity mode of glasses from series A, and glass B-1 from series B, overlaid for better comparison.

Figure 11 : ²³ Na MAS NMR spectra of series B (Example 2).

Figure 12: X-ray diffractrogram of glass sample B-2 showing the halo from the glassy phase and sharp peaks from unreacted S1O2. No crystalline molybdate phases are observed.

Figure 13: ¹¹ B MAS NMR spectra of series B (Example 2).

Figure 14: ²³ Na MAS NMR spectra of series C (Example 3).

Figure 15: X-ray diffractogram of glass sample C-2 which is devoid of reflection.

Figure 16: ¹¹ B MAS NMR spectra of series C.

Figure 17: ³¹ P MAS NMR spectra of samples A-3 (Example-1) and C-2

(Example-3).

DESCRIPTION OF THE PREFERRED EMBODIMENTS Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the invention belongs. Although any methods and materials similar or equivalent to those described herein can be used in the practice or testing of the present invention, the preferred methods and materials are now described. All publications mentioned hereunder are incorporated herein by reference.

High-level radioactive waste from spent-fuel reprocessing is immobilized in boroaluminosilicate glasses. However, some constituents of typical waste streams resist incorporation into the glass, instead forming crystallization products which can sequester radioactive ions from the glass, and are water soluble. For example, the low solubility of molybdenum and sulfur in widely used nuclear waste glasses restricts the overall amount of waste that can be loaded into a glass, consequently increasing glass volume (and hence repository size) and energy usage (from high-temperature melting).

The current strategy for preventing devitrification of radioactive ions is to restrict waste loading to well below the known solubility limits of the constituent waste products. While some exotic glasses and ceramics have been designed with very high waste loading capacities for certain ions, they are generally not very versatile (i.e. , they don't work well for other ions) and often require radical changes to production facilities.

As discussed herein, phosphorus, in the form of sodium phosphate, is included in the batch composition of model nuclear waste glasses based on the widely accepted aluminoborosilicate glass SON68, replacing up to 10 mol% of the silica in the base glass without any other compositional changes.

According to an aspect of the invention, there is provided a modified borosilicate glass, for example, an aluminoborosilicate glass, wherein or characterized in that up to 10 mol% of the silica has been replaced with P2O5. In some embodiments, the borosilicate glass is similar to the French reference glass, SON68, but the invention does not require this to be so. As will be appreciated by one of skill in the art, such a borosilicate glass or alumioborosilicate glass can be used for improved nuclear waste disposal, as discussed herein.

The main reason phosphate and silicate glasses have not been widely explored for nuclear waste disposal is that the network structure of glasses made of roughly equal amounts of phosphate and silicate is very complicated and prone to phase separation. A small amount of silica added to a predominantly phosphate glass results in unusual structural features in the silicate parts which nullify its ability to bond with radioactive waste ions. When a small amount of phosphate is added to a predominantly silicate glass, soluble orthophosphate units are formed, which reduces the chemical durability. As such, mixing phosphate and silicate glasses will bring out the worst properties in each.

In addition, phosphate is corrosive, and that is one reason it is not preferred. However, there is a small amount of phosphate present in many waste streams (<0.5%) which does not corrode melters or canisters. For example, other glass projects in our lab with higher P levels have resulted in corrosion.

However, as discussed herein, adding a small amount of phosphate to a boroaluminosilicate glass, the boron and aluminum in the glass prevent nuc!eation of phosphates, interconnect isolated phosphates (rendering them more tightly bound to the glass network), while retaining the high durability of silicates. The role of aluminum and boron in the silica-rich glass gets around the obvious problems associated with simply mixing silica with phosphate.

Thus, as discussed above, phosphorus is popularly thought to produce glasses with poor chemical durability and has rarely been considered a candidate for improving long-term performance of silicate glasses. Phosphorus also tends to phase separate from silicate-based glass networks, which further degrades the chemical durability. However, by using small amounts of phosphorus in the presence of a robust, multicomponent oxide network consisting of aluminum, boron and silicon, the chemical durability is not substantially reduced and crystallization is prevented by the inherent disorder in the complex network structure. Relatively well dispersed phosphate units - supported within the dominant aluminoborosilicate network - provide effective binding sites for molybdenum and sulfur, while remaining sufficiently isolated to prevent clustering, nucleation and crystallization. Thus, the addition of phosphorus to nuclear waste glasses increases the amount of molybdenum and sulfur that can be incorporated into the glass structure without crystallization of water-soluble molybdates and sulfates, thereby removing two significant limitations on waste loading capacities.

The amount of Mo and S that can be added to the base glasses without causing devitrification (i.e., formation of crystalline molybdate and/or sulfate phases) increases by several times with respect to the P-free versions of these glasses, which are similar to those used industrially, as discussed herein. Integration of Mo and S into the glassy network prevents the loss of radionuclides such as ¹³⁵ Cs, ¹³⁷ Cs and ⁹⁰ Sr through leaching into the environment via their sequestration into crystalline products, thereby facilitating long-term durability. The incorporation of Mo into the glass does not appear to reduce the simultaneous incorporation of S into the glass, indicating that this material is sufficiently versatile to immobilize multicomponent wastestreams.

Specifically, the incorporation of molybdenum and sulfur in aluminoborosilicate glasses was studied using solid-state nuclear magnetic resonance (NMR)

spectroscopy and X-ray diffraction (XRD), as discussed herein. The base

aluminoborosilicate glass is incapable of retaining Mo and S above the level of about 1 mol%, producing crystalline alkali molybdates and sulfates identified and quantified by NMR and XRD. The extent of molybdenum solubility was significantly improved by substituting phosphate for a small fraction of the silicate glass-network-formers, increasing Mo incorporation by a factor of three with 5 mol% P2O5, and to nearly five times the P-free glass solubility limit with the addition of 10 mol% P2O5. At the 5 mol% P2O5 additive level, both molybdenum and sulfur solubilities were simultaneously increased relative to the P-free glass, implying that their respective glass binding sites are independent. NMR measurements of the homogeneous Mo- and S-loaded glasses suggest that the phosphate network connectivity is key to their enhanced solubilities. As discussed herein, the substitution of phosphate for a small amount of silicate network formers in model glasses closely related to accepted compositions for radioactive waste disposal results in a significantly enhanced capacity for Mo and S incorporation. NMR spectroscopy, supported by x-ray diffraction, reveals that only at the lowest P concentrations do crystalline molybdates and sulfates form. The solubility of Mo is increased at least four-fold with respect to the P-free aluminoborosilicate nuclear glasses currently in use worldwide. Sulfur solubility is also substantially improved, even in the presence of Mo, suggesting that different binding sites are involved in vitrification. NMR spectroscopy of boron and phosphorus provides further structural clues about the specific bonding locations of Mo and S within the glassy network. The basis for this effectiveness is proposed to reside in the high cation-field- strength of P ⁵⁺ in the glass network, which can compete with Mo ⁶⁺ and S ⁶⁺ for nearby oxygens, thereby depriving the latter of ordered nucleation sites and preventing crystallization and phase separation. With this demonstration of the value of adding phosphate to the traditional nuclear waste glass compositions, higher waste-loading levels are accessible without compromising long-term performance.

According to an aspect of the invention, there is provided a composition for the preparation of borosilicate glass comprising:

30-60 mol% S1O2;

10-30 mol% B2O3;

2-30 mol% P2O5;

0-15 mol% AI2O3;

10-35 mol% Alkali oxide or alkaline earth oxide;

0-10 mol% M0O3; and

0-10 mol% SOs.

In another aspect of the invention, there is provided a composition for the preparation of borosilicate glass comprising:

45-55 mol% S1O2;

15-20 mol% B2O3;

2-10 mol% P2O5; 0-5 mol% AI2O3;

15-20 mol% Alkali oxide or alkaline earth oxide;

0-5 mol% M0O3; and

0-5 mol% SOs.

According to an aspect of the invention, there is provided a composition for the preparation of borosilicate glass comprising:

30-60 mol% S1O2;

10-30 mol% B203;

2-30 mol% P2O5;

0-15 mol% AI2O3; and

10-35 mol% Alkali oxide or alkaline earth oxide.

In another aspect of the invention, there is provided a composition for the preparation of borosilicate glass comprising:

45-55 mol% S1O2;

15-20 mol% B2O3;

2-10 mol% P2O5;

0-5 mol% AI2O3; and

15-20 mol% Alkali oxide or alkaline earth oxide.

In some embodiments, the alkali oxide is selected from the group consisting of: U2O, Na20, K2O, Rb20, CS2O and mixtures thereof.

In other embodiments, the alkali oxide is replaced fully or in part by alkaline earth oxides selected from the group consisting of: MgO, CaO, SrO, BaO and mixtures thereof.

Accordingly, in some embodiments of the invention, the alkali oxide or alkaline earth oxide is selected from the group consisting of: L12O, Na20, K2O, Rb20, CS2O, MgO, CaO, SrO, BaO and mixtures thereof.

As will be appreciated by one of skill in the art, in some embodiments, the base glass comprises S1O2, B2O3, AI2O3, Na20 and P2O5 wherein Na20 may be replaced partially or entirely by other alkali oxides or alkaline-earth oxides, as discussed above. As can be seen above, M0O3 and SO3 are shown as optional components. In some embodiments, these compounds as well as CS2O may be added to the composition to simulate the waste species. However, in use, these compounds could be replaced by many types of other oxides which are present in typical radioactive waste streams, for example but by no means limited to oxides of Zn (for example ZnO), Cr (for example Cr203), Sn, Fe, La, Nd, Cd, Y, Ag, Ni, Pr.

The nuclear fuel used in the reactors is in the form of UO2 pellets which are stacked in Mg-AI claddings. Once the fuel is spent, it is removed from the reactor core and either stored in repositories as such or further processed. The processing is what generates the nuclear waste which has to be immobilized in glass. The spent fuel is processed to extract left over fissile materials like ²³⁵ U, ²³⁹ Pu and fertile materials like ²³⁸ U. The process is termed as PUREX process and involves lots of organic solvents and inorganic acids. The slurry left over after extracting fissile and fertile materials will be rich in fission products which are highly radioactive in nature. This slurry is stored in interim tanks inside the reprocessing plants until the fission products with short half- lives have decayed. This slurry is calcined and converted into a powder which contains all the fission products in an oxide form.

There are two ways the calcined powder can be immobilized in glass:

a) The calcined powder is mixed with pre-made glass frits of fixed composition, melted at ~ 1 100°C and poured into stainless steel canisters.

b) Oxides used in glass melting (S1O2, B2O3, Na20, AI2O3, P2O5) are added directly to the calcined powder in fixed ratios and melted for sufficient time at ~1 100°C and poured into stainless steel canisters. As will be appreciated by one of skill in the art, this method has the distinct advantage that the ratio of glass to calcined waste can be freely altered and the glass composition can be easily customized if there are any variations in the compositions of calcined waste.

The calcined waste generated by this process varies in composition between different types of reactors used and different indigenous reprocessing methods followed. However, the specific chemical composition of such wastes are well known in the art. The amount of nuclear waste immobilized in the glass is limited to the solubility limit of the least soluble fission products. High-field-strength fission products such as Mo exhibit very low solubility. Sulfate (S ⁶⁺ ) added during reprocessing is also sparingly soluble in the borosilicate glass. The current nuclear waste loading capacity in borosilicate glasses is 12 wt%, which corresponds to 1 mol% Mo and 0.5 mol% S immobilized in glass without any phase separation.

When the glass is loaded with nuclear waste (up to 12 wt%), certain fission products are integrated directly into the glass network while some fission products are associated with the borosilicate network, distributed randomly as isolated structural units. The main reason to limit the amount of nuclear waste added into the glass is to avoid any sort of phase separation and eventual crystallization. Above the solubility limit of the least-soluble fission product, for example Mo, radioactive isotopes are sequestered from the glass and phase separate from the glass as alkali-molybdate or alkaline earth molybdate.

Pure silicate glasses are not good candidates for nuclear waste immobilization as they have very high processing temperatures. However, these glasses are extremely durable. On the other hand, phosphate glasses can be melted at low temperatures and can immobilize very high amounts of nuclear waste. However, these glasses exhibit poor chemical durability. As discussed herein and as known by those of skill in the art, phosphate glasses are capable of immobilizing up to 50 wt% of nuclear waste.

As discussed above, adding a small amount of phosphorus to a silicate glass can have both positive and negative effects. For example, adding phosphorus might compromise the glass durability and physical properties, but positive effects such as low melting temperatures and high waste loading capacities are achieved.

As discussed herein, we have observed a three-fold increase in Mo and S solubility when 5 mol% of S1O2 is replaced with P2O5 in a borosilicate glass. This corresponds to 30-40 wt% of nuclear waste immobilization, considering a realistic multicomponent waste stream. By this approach a three-fold reduction in the volume of glass should be achievable The strategy of using a small amount (<10 mol%) of a high-field-strength network-forming cation (P ⁵⁺ ) to compete with minor high-field-strength cations and interfere with crystallization tendencies is likely to be adaptable to many other low- solubility radioactive waste products. Lanthanides and platinoids, along with a wide variety of highly charged transition metals, are often present in complex radioactive wastes, and should be amenable to immobilization by the use of phosphorus additives. Other high-field-strength network-forming cations (e.g., Ge ⁴⁺ , Ti ⁴⁺ , Ga ³⁺ , Zr ⁴⁺ ) may also prove valuable as additives to aluminoborosilicate glasses to

devitrification.

The invention will now be further elucidated and explained by way of examples; however, the invention is not necessarily limited to the examples.

Example 1 : Series A - the role of phosphorus in retaining Mo in the glass

²³ Na MAS NMR of Example 1 shows that the boroaluminosilicate glass alone, without phosphorus, cannot contain 3 mol% Mo in the glass (Figure 1 , A-1). The sharp peaks between 0 to -10 ppm in the glass sample A-1 are indicative of crystalline sodium molybdate phases (Na2Mo0 ₄ and Na ₂ Mo0 ₄ .2H20) [36], as shown in the fitted spectrum, Figure 2. Specifically, as will be appreciated by one of skill in the art, crystalline phases are observed in samples A-1 and A-2 whereas samples A-3 and A- 4 are homogeneous glasses without any phase separation, as shown in Figure 1. The NMR data for this sample indicate that 4% of the Na ⁺ ions are present in crystalline molybdate phases, representing a substantial fraction of the Mo added to the glass. While ²³ Na NMR can only detect the chemical environment of sodium ions, this result is supported by the indexed sodium molybdate reflections observed by x-ray diffraction (Figure 3). That is, the reflections in the X-ray diffractogram of sample A-1 are assigned to sodium molybdate and sodium molybdate dehydrate crystalline phases. It is worth noting that the reflections are observed from both sodium molybdate and hydrated sodium molybdate phases, reflecting the affinity of alkali molybdates for atmospheric water. Despite subjecting the devitrified glass sample to x-ray diffraction analysis immediately after grinding the sample on the benchtop, the anhydrous sodium molybdate absorbed water from the air and partially transformed into the hydrated form. This extremely hygroscopic behaviour underscores the importance of preventing molybdate crystallization in materials intended for long-term nuclear waste immobilization. Specifically, the resulting crystalline phases are highly water-soluble and will leach out of the material into the environment, taking radionuclides with them.

With the addition of 2.5 mol% P2O5 (A-2), almost all of the Na is represented by the broad featureless ²³ Na NMR peak indicative of sodium in a glassy environment, but there is a hint of crystallinity (<0.5%) in the same region as A-1 (Figure 1).

Glasses with 5.0 and 7.5 mol% P2O5 show no evidence of crystallinity, indicating that all Na is present in purely glassy environments. X-ray diffraction provides no evidence of crystalline phases in these glasses.

¹¹ B MAS NMR of this series with increasing phosphate provides insight into how Mo is incorporated into the glassy network. The ¹¹ B MAS NMR spectra of glasses in series A are shown in Figure 4; the same spectra are presented in absolute-value mode and overlaid in Figure 5 for a more direct comparison. The spectra consist of two main peaks: three-coordinate boron ( ^[3] B) appears between 10 and 20 ppm, and four-coordinate boron ( ^[4] B) appears between -2 and 5 ppm [29] Fits obtained from spectral deconvolution with three ^[3] B sites and two ^[4I B sites are shown in Figure 6 and their associated parameters are listed in Table 3. Note that whereas the two ^t4] B sites used to fit the data represent structurally distinct tetrahedral boron units (see below), three components were used to fit the ^[3] B region only to account for its total intensity, and does not imply the presence of distinct trigonal borons. With increasing phosphate, the fraction of tetrahedral borate units decreases, quantified by N ₄ = ( ^[4 ]B/[4]B+[3]B) _I Figure 7. A related observation is that the two types of tetrahedral boron making up the composite ^t4] B peak around 2 ppm change in relative intensity, as shown in Figure 8.

As such, as will be appreciated by one of skill in the art, Figure 4, 5, 6, 7 and 8 provide information on boron coordination and its chemical environment. It is found that the amount of tetrahedral boron in the glass decreases as phosphorus replaces Si. Figure 8 provides quantitative information about boron bonding to neighbouring silicate and phosphate units.

To understand these changes, it is necessary to identify the possible structures of the four-coordinate borons. In alkali borosilicate glasses, peaks at -1 and -2 ppm have been attributed to four-coordinated boron bonded to three (MB3S ₁ ) and four Si units ( ^[4] B ₄ si), respectively [29] A ^[4] B unit in a borophosphate glass network bonded to four neighboring phosphate units ( ^[4] B _4p ) has a chemical shift of ~ -4 ppm, whereas ^[4] B units bonded to fewer than four phosphate units (e.g., ^[4] B3P, where the remaining linkage is to trigonal boron) have chemical shifts ~ -1.8 ppm [37] However in a borosilicate glass, phosphate units are distributed as isolated tetrahedral units

(orthophosphates) not directly bonded to the extended borosilicate network [38], hence it is highly likely that the two ^[4] B peaks in these spectra arise from boron units bonded to three and four silicate tetrahedra.

In a borophosphate glasses with high Mo concentrations (>30 mol%), Mo ⁶⁺ ions are known to bond to ^[4] B through Mo-O-B linkages [37,39] However, in an aluminoborosilicate glass, molybdate anions (Mo0 ₄ ^2‘ ) are concentrated in alkali- and alkaline-earth-rich domains within the glass and do not connect directly to the aluminoborosilicate glass network [40] When P2O5 is substituted for S1O2, the number of Si units surrounding ^[4] B units decreases, resulting in a reduction of the ^[4] B ₄ si peak ( ^[4 lB-2) at ~0 ppm. Tetrahedral borons which have fewer than four neighboring Si units are either transformed into ^[4] B3Si or ^[3] B units. This explains both why the intensity of WBssi changes very little throughout the series ( ^[41 B-1), and the intensity of ^[3] B units gradually increases, as shown in Figure 4.

³¹ P MAS NMR spectra of Mo-containing glasses with increasing P content show a progressive peak shift to lower frequency (Figure 9), which indicates a structural change toward a greater degree of phosphate chain connectivity [41].

Whereas the 2.5 mol% glass shows predominantly isolated phosphate units

(orthophosphates, +3 ppm), doubling the phosphate content significantly increases the fraction of phosphate dimers (pyrophosphates, -5 ppm) and longer phosphate chains (metaphosphates, -10 ppm). Increasing the phosphate content even further to 7.5 and 10 mol% - along with a small increase in Mo content - converts even more of the phosphorus to metaphosphate chains. An overlay of all the spectra from the series clearly highlights the overall transformation of phosphate network with a gradual increase in the phosphorus content of the glass (Figure 10). Since it is known that chain phosphate oxygens are more likely to bond Mo than borate oxygens in low- Mo glasses [39], the increase in such units helps explain why Mo is more effectively incorporated into the glassy phosphate network, rather than clustering into (pre)- crystalline molybdate-rich regions. That is, Figures 9 and 10 explain how P is integrated into the borosilicate network, where increased phosphorus content phosphates form chains.

Example 2: Series B - how much molybdenum can phosphate-containing glasses hold?

An aluminoborosilicate glass with added phosphorus is capable of

incorporating much higher levels of Mo than a pure aluminoborosilicate glass. Series A proves that with 7.5 mol% P2O5, more than three times as much Mo can be immobilized in the glassy phase. In series B, the Mo content is raised to 5 mol% in a glass containing 10 mol% P2O5 to test the limits of phosphorus as a crystallization- prevention additive. Neither ²³ Na MAS NMR (Figure 1 1) nor x-ray diffraction (Figure 12) show any evidence of molybdate crystallization in these glasses, despite the presence of nearly five times the Mo solubility limit in boroaluminosilicate glasses [17] Specifically, Figure d 1 shows that there are no sodium-bearing crystalline phases in these glasses, even though they are loaded with 4 mol% (B-1 ) and 5 mol% M0O3 (B- 2) at a P2O5 concentration of 10 mol%. Furthermore, the X-ray diffractogram of Figure 12 lacks the reflections from sodium molybdate phases which supports the NMR results that the glasses with high Mo contents are homogeneous. It should be noted that XRD reveals the presence of a small amount of unreacted S1O2 in glass B-2. While this crystalline impurity is unsightly, it underscores the efficacy of P to inhibit molybdate formation, even when other crystalline phases are present to serve as nucleation sites. The ¹¹ B MAS NMR spectra in Figure 13 are very similar to each other, and to sample A-4 (Figure 4), indicating that with a sufficiently high amount of phosphorus, small increases in the amount of Mo (i.e. , from 4 to 5 mol%) have little influence on the borate portion of the glass network. Likewise, ³¹ P MAS NMR spectra are almost indistinguishable in series B (not shown), implying that the glass network structure is little changed by the addition of more Mo over this range. It may be reasonably concluded that there are ample Mo binding sites available along the phosphate chains without alteration of the degree of polymerization.

Example 3: Series C - the incorporation of sulfur into P-containing glasses

An aluminoborosilicate glass without phosphorus or molybdenum charged with 3 moi% SO3 (C-1) yields a strong sodium sulfate signal in the ²³ Na MAS NMR spectrum (Figure 14), highlighting the poor sulfate solubility of conventional nuclear waste glasses. By contrast, a comparison of the ³¹ Na MAS NMR spectra of A-3 (Figure 1) and C-2 (Figure 14) shows the effect of adding SO3 to a 5 mol% P ₂ O5 glass which already contains 3 mol% M0O3: neither ²³ Na MAS NMR (Figure 14, C-2) nor XRD (Figure 15) shows any evidence of crystallinity, despite the fact that the glass is already loaded with Mo. Specifically, the sodium NMR spectra of Figure 14 shows crystalline sodium sulfate (sharp peak) in P-free glasses and no sodium sulfate in P- containing glass, where all sodium is in the glassy phase (broad peak). Similarly, the X-ray diffractogram of Figure 15 lacks the reflections from sodium molybdate or sodium sulfate crystalline phases which supports the NMR results that the glass (C-2) can dissolve both Mo and S in 5 mol% P2O5 glass.

Since sodium sulfate is a water-soluble phase known to incorporate other alkali metals which could be radioactive, it is important to avoid the formation of this compound. Hence, phosphate - even at the fairly low concentration of 5 mol% - may be used to prevent sulfate crystallization and maintain sulfur within the glassy network.

¹¹ B MAS NMR of sulfate-containing glasses (Figure 16) signify relatively large changes in the borate network, indicating both a shift in the balance of tetrahedral and trigonal boron network species, and non-negligible frequency shifts with the addition of Mo and P2O5, similar to what is shown in Figures 4-8.

³¹ P MAS NMR spectra (Figure 17) are especially valuable for monitoring structural changes upon the addition of sulfate to an Mo-bearing 5 mol% P2O5 glass, showing a strong reduction in the chain phosphate signal around -7 ppm, indicative of reduction in the number of metaphosphates units. The reduction in the intensity of the metaphosphate units provides compelling evidence that P-O-P bonds are being broken and P-O-S bonds formed, explaining the effective incorporation of sulfur into the glass network as the formation of P-O-S bonds is crucial in the solubility of sulfur in this glass system.

Example 4 - Synthesis

Compositions of glasses from all three series are listed in Table 1. The glasses were synthesized by mixing S1O2, AI2O3, M0O3, Na2C03, CS2CO3, (NaP03)6, Na2S0 ₄ and B2O3 in 2-gram batches as per the series composition, de-carbonating at 650°C for 12 hours and melting at 1 100°C for one hour in Pt/Au crucibles with the lids on, before cooling at ca. 5°C/min. The glasses were ground and remelted at 1 100°C for 60 minutes to ensure homogeneous mixing of oxides and slow cooled again. B2O3 was synthesized by heating H3BO3 at 450°C, 650°C and 800°C for 30, 15 and 15 minutes respectively. All other reagents were used as received from commercial sources. The final glass samples were placed in tight-capped glass vials and stored in a desiccator prior to NMR and XRD analysis.

Example 5 - Characterization:

NMR analysis was carried out on a Varian 600 MHz NMR spectrometer equipped with a Chemagnetics 1.6 mm triple-resonance solid-state magic-angle- spinning (MAS) NMR probe. All samples were spun at the magic angle with a spinning speed of 30 kHz (±5 Hz). At a magnetic field of 14.08 T, the resonance frequencies of ¹¹ B, ²³ Na, and ³¹ P are 192.54, 158.74, and 242.86 MHz, respectively. Bloch-decay experiments were carried out with pulse lengths and relaxation delays, respectively, of 0.35 ps and 2 s ( ¹¹ B), 0.39 ps and 4 s ( ²³ Na), 1 ps and 30 s ( ³¹ P). For ¹¹ B and ²³ Na, these pulse lengths corresponded to a tip angle of 15°, which will equally excite sites with different quadrupolar coupling constants (Co). For ³¹ P, the tip angle was 30° to maximize signal intensity for a given acquisition time. The MAS NMR signal was averaged over 1024, 1024 and 128 transients for ¹¹ B, ²³ Na and ³¹ P isotopes, respectively. ²³ Na, ¹¹ B and ³¹ P MAS NMR spectra were referenced using aqueous solutions of 0.1 M NaCI at 0 ppm, 0.1 M FI3BO3 at +19.6 with respect to trifluoroboroetherate, and 85% H3PO4 (0 ppm), respectively.

Spectral fitting was carried out using the DMFIT software developed by Massiot et. al. [33] In the case of ¹¹ B MAS NMR spectra, the four-coordinate borons ( ^[4] B) were fit with a Gaussian/Lorentzian peakshape, and the three-coordinated borons ( ^[3] B) were fit with Quasar [34] to better model the quadrupolar lineshape. The ^[4] B peak integration was adjusted for overlapping satellite transition intensity by

subtracting the latter integral area from the former. For ²³ Na MAS NMR spectra, Quasar was used to fit the sharp peaks in the spectrum representing

crystalline phases, and a Gaussian/Lorentzian lineshape was used to fit the broad peak associated with sodium in the glassy phase.

X-ray diffraction analysis was carried out using a D4 Endeavor Bruker X-ray diffractometer. The samples were finely ground and mounted on polymer diffraction plates and subjected to Cu-Ka radiation generated at voltage and current of 40 kV and 40 mA, respectively. A 2Q range of 10° and 50° was scanned with a step size of 0.02° and an acquisition time of 30 seconds, corresponding to 2000 steps. The sample was spun at 15 rpm during data acquisition to ensure homogeneous sampling. Phase identification was done using the built-in matching software to search the ICDD [35], and all combinations of elemental compositions were

considered.

While the preferred embodiments of the invention have been described above, it will be recognized and understood that various modifications may be made therein, and the appended claims are intended to cover all such modifications which may fall within the spirit and scope of the invention. REFERENCES

[1] Donald I.W. et. al, Review The immobilization of high level radioactive wastes using ceramics and glasses, J. Mater. Sci. 32 (1997) 5851-5887.

[2] B.J. Greer, S. Kroeker, Characterisation of heterogeneous molybdate and chromate phase assemblages in model nuclear waste glasses by multinuclear magnetic resonance spectroscopy, Phys. Chem. Chem. Phys. 14 (2012) 7375.

[3] N. Chouard, D. Caurant, O. Majerus, J.L. Dussossoy, S. Klimin, D. Pytalev, R.

Baddour-Hadjean, J.P. Pereira-Ramos, Effect of M0O3, Nd2C>3, and Ru0 ₂ on the crystallization of soda-lime aluminoborosilicate glasses, J. Mater. Sci. 50 (2015) 219- 241.

[4] D. Caurant, O. Majerus, E. Fadel, A. Quintas, C. Gervais, T. Charpentier, D. Neuville, Structural investigations of borosilicate glasses containing M0O3 by MAS NMR and Raman spectroscopies, J. Nucl. Mater. 396 (2010) 94-101.

[5] S. Kroeker, S. Schuller, J.E.C. Wren, B.J. Greer, A. Mesbah, ¹³³ Cs and ²³ Na MAS NMR Spectroscopy of Molybdate Crystallization in Model Nuclear Glasses, J. Am. Ceram. Soc. 99 (2016) 1557-1564.

[6] J.D. Vienna, Nuclear Waste Vitrification in the United States: Recent Developments and Future Options, Int. J. Appl. Glas. Sci. 1 (2010) 309-321.

[7] D. Manara, A. Grandjean, O. Pinet, J.L. Dussossoy, D.R. Neuville, Sulfur behavior in silicate glasses and melts: Implications for sulfate incorporation in nuclear waste glasses as a function of alkali cation and V2O5 content, J. Non. Cryst. Solids. 353 (2007) 12-23.

[8] A.L. Billings, K.M. Fox, Retention of Sulfate in Savannah River Site High-Level Radioactive Waste Glass, Int. J. Appl. Glas. Sci. 1 (2010) 388-400.

[9] L. Wu, J. Xiao, X. Wang, Y. Teng, Y. Li, Q. Liao, Crystalline phase, microstructure, and aqueous stability of zirconolite-barium borosilicate glass-ceramics of immobilization of sulfate bearing high-level liquid waste, J. Nucl. Mater. 498 (2018) 241-248.

[10] C.P. Kaushik, R.K. Mishra, P. Sengupta, A. Kumar, D. Das, G.B. Kale, K. Raj, Barium borosilicate glass - a potential matrix for immobilization of sulfate bearing high-level radioactive liquid waste, J. Nucl. Mater. 358 (2006) 129-138.

[1 1] P.A. Bingham, S. Vaishnav, S.D. Forder, A. Scrimshire, B. Jaganathan, J. Rohini, J.C.

Marra, K.M. Fox, E.M. Pierce, P. Workman, J.D. Vienna, Modelling the sulfate capacity of simulated radioactive waste borosilicate glasses, J. Alloys Compd. 695 (2017) 656- 667.

[12] D.A. Mckeown, I.S. Muller, H. Gan, I.L. Pegg, C.A. Kendziora, Raman studies of sulfur in borosilicate waste glasses: sulfate environments, J. Non. Cryst. Solids. 288 (2001) 191-199.

[13] B.J. Riley, J.S. McCloy, A. Goel, M. Liezers, M.J. Schweiger, J. Liu, C.P. Rodriguez, D.S. Kim, Crystallization of rhenium salts in a simulated low-activity waste borosilicate glass, J. Am. Ceram. Soc. 96 (2013) 1150-1 157.

[14] B. Brendebach, M.A. Denecke, G. Roth, S. Weisenburger, Sulfur incorporation in high level nuclear waste glass: A S K-edge XAFS investigation, J. Phys. Conf. Ser. 190 (2009).

[15] D. Caurant, O. Majerus, E. Fadel, A. Quintas, C. Gervais, T. Charpentier, D. Neuville, Structural investigations of borosilicate glasses containing Mo0 ₃ by MAS NMR and Raman spectroscopies, J. Nucl. Mater. 396 (2010) 94-101.

[16] D. Caurant, O. Majerus, E. Fadel, M. Lenoir, C. Gervais, O. Pinet, Effect of molybdenum on the structure and on the crystallization of Si02-Na20-Ca0-B203 glasses, J. Am. Ceram. Soc. 90 (2007) 774-783.

[17] S. Kroeker, S. Schuller, J.E.C. Wren, B.J. Greer, A. Mesbah, ¹³³ Cs and ²³ Na MAS NMR Spectroscopy of Molybdate Crystallization in Model Nuclear Glasses, J Am. Ceram. Soc. 99 (2016) 1557-1564.

[18] M. Lenoir, A. Grandjean, S. Poissonnet, D.R Neuville, Quantitation of sulfate solubility in borosilicate glasses using Raman spectroscopy, J. Non. Cryst. Solids. 355 (2009) 1468-1473.

[19] G. Tricot, K. Ben Tayeb, L. Koudelka, P. Mosner, H. Vezin, Insertion of M0O3 in borophosphate glasses investigated by magnetic resonance spectroscopies, J. Phys. Chem. C. 120 (2016) 9443-9452.

[20] P.A. Bingham, A.J. Connelly, R.J. Hand, N.C. Hyatt, P.A. Northrup, Incorporation and speciation of sulphur in glasses for waste immobilisation, Glas. Technol. Eur. J. Glas. Sci. Technol. Part A. 50 (2009) 135-138.

[21] P.A. Bingham, R.J. Hand, Sulphate incorporation and glass formation in phosphate systems for nuclear and toxic waste immobilization, Mater. Res. Bull. 43 (2008) 1679- 1693.

[22] P.A. Bingham, R.J. Hand, S.D. Forder, A. Lavaysierre, S.H. Kilcoyne, I. Yasin, Preliminary studies of sulphate solubility and redox in 60 P ₂ O ₅ -40 Fe ₂ Q ₃ glasses, Mater. Lett. 60 (2006) 844-847.

[23] M. Szumera, I. Wacawska, Effect of molybdenum addition on the thermal properties of silicate-phosphate glasses, J. Therm. Anal. Calorim. 109 (2012) 649-655.

[24] S. Schuller, F. Bart, Phase separation and crystallization induced by the addition of Mo and P to a calcium-sodium silicate glass, Glas. Odyssey 6th ESG Conf. (2002).

[25] S. Schuller, O. Pinet, B. Penelon, Liquid-liquid phase separation process in borosilicate liquids enriched in molybdenum and phosphorus oxides, J. Am. Ceram. Soc. 94 (201 1 ) 447-454.

[26] S. Schuller, O. Pinet, A. Grandjean, T. Blisson, Phase separation and crystallization of borosilicate glass enriched in Mo03, P205, Zr02, CaO, J. Non. Cryst. Solids. 354 (2008) 296-300.

[27] O. Pinet, J.L. Dussossoy, C. David, C. Fillet, Glass matrices for immobilizing nuclear waste containing molybdenum and phosphorus, J. Nucl. Mater. 377 (2008) 307-312.

[28] K.B. Patel, B. Boizot, S.P. Facq, S. Peuget, S. Schuller, I. Farnan, Impacts of composition and beta irradiation on phase separation in multiphase amorphous calcium borosilicates, J. Non. Cryst. Solids. 473 (2017) 1-16.

[29] C. Martineau, V.K. Michaelis, S. Schuller, S. Kroeker, Liquid-liquid phase separation in model nuclear waste glasses: A solid-state double-resonance NMR study, Chem. Mater. 22 (2010) 4896-4903.

[30] L.M. Funke, H. Eckert, Charge Compensation in Sodium Borophosphate Glasses Studied by ¹¹ B{ ²³ Na} and ³¹ P{ ²³ Na} Rotational Echo Double Resonance Spectroscopy, J. Phys. Chem. C. 120 (2016) 3196-3205. doi:10.1021/acs.jpcc.5b1 1608.

[31 ] S. Sen, P. Yu, V.P. Klyuev, B.Z. Pevzner, Atomic structure of beryllium boroaluminate glasses: A multi-nuclear NMR spectroscopic study, J. Non. Cryst. Solids. 354 (2008) 4005-401 1.

[32] R. Bouakkaz, A. Abdelouas, Y. El Mendili, B. Grambow, S. Gin, SON68 glass alteration under Si-rich solutions at low temperature (35-90°C): kinetics, secondary phases and isotopic exchange studies, RSC Adv. 6 (2016) 72616-72633.

[33] D. Massiot, F. Fayon, M. Capron, I. King, S. Le Calve, B. Alonso, J.O. Durand, B.

Bujoli, Z. Gan, G. Hoatson, Modelling one- and two-dimensional solid-state NMR spectra, Magn. Reson. Chem. 40 (2002) 70-76.

[34] J.P. Amoureux, C. Fernandez, QUASAR— Solid-State NMR Simulation for Quadrupolar Nuclei, Univ. Lille, Lille, Fr. (2005). [35] M. Hellenbrandt, The inorganic crystal structure database (ICSD) - Present and future, Crystallogr. Rev. 10 (2004) 17-22.

[36] J.E.C. Wren, B.J. Greer, V.K. Michaelis, C.S. Higman, S. Kroeker, Multinuclear Magnetic Resonance Investigation of Crystalline Alkali Molybdates, Inorg. Chem. 54 (2015) 9853-9861.

[37] J. Subcik, L. Koudelka, P. Mosner, L. Montagne, B. Revel, I. Gregora, Structure and properties of Mo03-containing zinc borophosphate glasses, J. Non. Cryst. Solids. 355 (2009) 970-975.

[38] Y. Yu, B. Stevensson, M. Eden, Medium-Range structural organization of phosphorus bearing borosilicate glasses revealed by advanced solid-state NMR Experiments and MD Simulations: Consequences of B/Si Substitutions, J. Phys. Chem. B. 121 (2017) 9737-9752.

[39] G. Tricot, K. Ben Tayeb, L. Koudelka, P. Mosner, H. Vezin, Insertion of M0O ₃ in borophosphate glasses investigated by magnetic resonance spectroscopies, J. Phys. Chem. C. 120 (2016) 9443-9452.

[40] A. Brehault, D. Patil, H. Kamat, R.E. Youngman, L.M. Thirion, J.C. Mauro, C.L. Corkhill, J.S. McCloy, A. Goel, Compositional Dependence of Solubility/Retention of Molybdenum Oxides in Aluminoborosilicate-Based Model Nuclear Waste Glasses, J. Phys. Chem. B. 122 (2018) 1714-1729.

[41] R.K. Brow, R.J. Kirkpatrick, G.L. Turner, Local Structure of xAI ₂ 0 ₃ .(1-x)NaP0 ₃ glasses:

An NMR and XPS study, J. Am. Ceram. Soc. 73 (1990) 2293-2300.