Ceramic and mineral waste forms

The concept of immobilizing the radioactive elements of nuclear waste in an assemblage of mineral phases was originally introduced by Hatch [1] at Brookhaven National Laboratory in 1953 (Tables 6.7 and 6.8). The feasibil­ity of making a ceramic of natural mineralogically stable phases was dem­onstrated by McCarthy [53, 54] and Roy [2, 55] at the Pennsylvania State University between 1973 and 1976. Since that time, a number of other mineralogic-ceramic assemblages have been developed. Among these high temperature (1000-1500°C) processes are the Sandia titanate-based ceramic [56] [ the Australian titanate-based ceramic ‘Synroc’ [57-59] [ the silicate — phosphate supercalcine ceramics [60], the alumina-based tailored ceramics [61,62], and the Pu pyrochlores [63]. Often in ceramics made by cold press­ing and sintering or hot isostatic pressing, an intergranular glassy phase is produced during liquid phase sintering on the ceramic grain boundaries and the radionuclides preferentially migrate to the glassy phase(s) [64-72] . If the radionuclides are incorporated in the intergranular glassy phase(s), they have been found to leach at about the same rates as those from glassy waste forms [22] .

Crystalline (ceramic/mineral) waste forms made by moderate tempera­ture (700-750°C) thermal treatment have not been as intensely investigated as those formed at high temperatures as discussed above [61] . However, crystalline wasteforms made from clay have also been studied almost con­tinuously since the work of Hatch in 1953 [1,61]. Often the high temperatures used for sintering of supercalcine ceramics created sodalite-cancrinite mineral assemblages. In 1981, Roy [73] proposed low temperature hydro­thermally processed low solubility phase assemblages consisting of the micas, apatite, pollucite, sodalite-cancrinite, and nepheline, many of which could be made from reaction of various clays (kaolin, bentonite, illite) with waste.

Clay-based crystalline (ceramic/mineral) waste forms were not pursued in the late 1970s and early 1980s because there was no continuous com­mercial technology available that could process the waste/clay mixtures in a hydrothermal environment [ 61] . A commercial facility to continuously process radioactive wastes by pyrolysis at moderate temperatures in a hydrothermal steam environment was built by Studsvik in Erwin, Tennessee in 1999 [74, 75] . This facility uses a fluidized bed steam reforming (FBSR) technology to pyrolyze 137Cs and 60Co organic resins from commercial nuclear facilities. This technology has the capability to process a wide variety of solid and liquid streams including wastes containing organic ion exchange resins, charcoal, graphite, sludge, oils, solvents, and cleaning solutions at radiation levels of up to 400 R/hr. When clay is added as a mineralizing agent, the feldspathoid minerals (sodalite, nosean and nepheline) are formed by nanoscale reaction with the clay. The phases formed act as hosts for high Cl, I, F, 99Tc, and SO4 alkali (Na, K, Cs) bearing wastes [76-80] and organics are destroyed creating steam and CO2. The mineralization occurs at the moderate FBSR temperatures because the FBSR operating tempera­ture is in the range in which most clays become amorphous at the nanoscale level, e. g. kaolin, bentonite (montmorillonite), and illite. The clays lose their hydroxyl (OH.) groups at the FBSR temperatures which destabilizes the octahedral (six nearest neighboring atoms that form an octagon) Al3+ cation in their structure (Fig. 6.2) and they become amorphous as confirmed by X-ray diffraction (XRD) analysis. The alkali in the waste ‘alkali activates’ the unstable Al3 + cation to form new mineral phases and the fluidizing agent, steam, catalyzes the mineralization. In the absence of steam, many of these mineral phases only form at temperatures of >1200°C.



6.2 Atomic structures of various clays (kaolin, bentonite — montmorillonite, illite). After [400, 401].

Ceramic waste forms can be single phase, e. g. UO2, or single-phase solid solutions like (U, Th, Pu)O2 (Table 6.7). Multiphase ceramics are formulated so that each radionuclide can substitute on a given host lattice in the various phases (see Table 6.8).

Of great importance when relying on the LRO (size and coordination of the crystallographic site which will act as host to a given radionuclide or its decay product upon transmutation) is that the crystal-chemical substitu­tions must be electrically balanced [ 81, 82] . When a monovalent cation transmutes to a divalent cation, the substitutions must be coupled to retain the electrical balance of the host phase without destroying the integrity of the phase: the lattice site must be of suitable size and bond coordination to accept the transmutation. The bonding in crystalline ceramic or mineral phases can only maintain charge balance in one of two ways: (1) if sufficient lattice vacancies exist or (2) if a variable valance cation like Fe or Ti is present in a neighboring lattice site for charge balance. Both scenarios assume that the variable valence cations do not change lattice sites and that
the charge balancing cations are in the same host phase in nearby lattice sites. The lattice site must be of sufficient size or flexible enough to accom­modate the transmuting cation. It is advantageous if the lattice site of the desired host phase has irregular coordination or is distorted as will be shown in some examples below.

The solubility or flexibility of a ceramic or mineral phase(s) as hosts for a substituted cation of a different valence can be studied by performing coupled substitutions on the phase pure mineral host phase. If the number of cations changes during the substitution, a vacancy is either created or consumed and the substitution must maintain electrical neutrality. These types of substitution are most often seen in polymorphic substitutions [83] of the type

□ + Ba2+ ^ 2K+

or □ + Ca2+ ^ 2Na+

or □ + Na+ + 2Ca2+ ^ 3Na+ + Ca2+

where □ denotes a vacancy. Implicit in these coupled substitutions is the fact that the exchanging cations occupy the same lattice sites, have the same coordination, and thus the crystallographic symmetry is maintained. There­fore, substitutions as described above should be written with roman numer­als that designate the number of oxygen atoms that coordinate around a given cation, e. g. VIIICa designates the octahedral VIII-fold coordination for the Ca2+ lattice site in oxyapatites:

3VIIICa2+, ^ 2VIIINd3+

host phase substituted phase

Calcium-neodymium coupled substitutions have been successful [81, 82] in the oxyapatite (Ca6[SiO4]3) structure forming completely substituted Nd4^2 [SiO4]3 where two-thirds of the lattice sites have Nd3+ and one-third are vacant. In the oxyapatite structure, the Ca3+ is normally in VIII-fold coor­dination and has a 1.12 A [84-86] atomic radius. The Nd3+ cation in VIII-fold coordination also has an atomic radius of 1.11 A [86] very close to the Ca2+ atomic radius in VIII-fold coordination. Felsche showed that the rare earth elements La3+ through Lu3+ can substitute for Ca3+ and form oxyapatites, RE4.67^0.33[SiO4]3O [87]; see Table 6.8]. McCarthy and Davidson [54] showed that even more complex, but coupled, substitutions were possible in the oxyapatite structure such as

2VIIICa2+ ^ 2.7VIIIINd3+ +1.7VIIIICs+ +0.86VIIIICe4+ + VIIII0.86Sr2++0.88^

host phase substituted phase

where the atomic radius, r, of Cs+ in VIII-fold coordination is 1.74 A, Ce4+ in VIII-fold coordination is 0.97 A, and Sr2+ in VIII-fold coordination is 1.26 A. In this case, small radii cations such as Ce4+ are mixed with large radii cations like Cs+ so that individual lattice sites can distort without perturbing the entire crystal structure. Note that the exchanging cations are always in the same lattice site of the same host phase [54, 81, 82, 87].

The substitutions such as given above for the oxyapatites were also dem­onstrated [ 81, 82] to be possible in many other Ca-bearing cementitious mineral phases such as larnite (Ca2SiO4 or P-C2S), alite (calcium trisilicate or Ca3SiO5 or C3S), C3A (Ca3Al2O6), and C4AF (Ca4Al2Fe2O10).This allowed Jantzen et al[ [88, 89] to make substitutions for Ca[+ in each phase (up to -15 mole%) and the following additional substitutions[19]:

Ca2+ + □ ^ 2Cs+

substituted phase

host phase

20 ^ Cs+ +Sr0.52+ +Nd3+0,17 +Ce4+0.25 +0.08d

host phase substituted phase

1.5Ca2++Si4+ ^ Sf2+ +Mo5+ +02

host phase substituted phase

4Ca2+ +Fe3+ +Al3+ ^ 0,66Nd3+ +Zr4+ +Mo4+ +Sr2+ +Ba2+ +1.33Ц

host phase substituted phase

42C2 + 22F23,

r~1.18 A r=0.65 A

host phase

^ 2.66IXNd+3 + 0.38VICe+4 + 0.56VIZr+4 + 0.75VIFe+3 +1.65^

2=1.16 A r= 2.87 A r =0.72 A r=0.65 A

The substitution is ordered upon fabrication and incommensurate super­structures result when Cs+ substitutes for Ba2+ [91].

Cesium has been experimentally substituted for Ba when Fe3+ is substi­tuted for Ti3+ in the VI-fold sites of hollandite. The species

VIIICs+0.28VIIIBa2+1.00 VIAl3+146VIFe3+o.82 VITi5.72O16 has been fabricated by

A site B site C site

sintering in air at 1320°C [92]. A Ba-Al hollandite (Ba116Al232Ti5 68O16) was electron irradiated (1-2.5 MeV) and P-irradiated (4 x 108 to7 x 109Gy) and found to contain Ti3+ centres and O2- superoxide ions which confirmed the mechanism of charge balance during transmutation [92]. Theoretically, the limiting y value in hollandite is 0.81 Cs which corresponds to a 9.54 wt% waste loading of Cs2O [93] .

Single-phase and multiphase ceramics can be made by many of the thermal treatment technologies given in Table 6.2. Examples include melting in smelters instead of melters, cold pressing (CIP or CUP) and sintering, hot isostatic pressing (HIP), or hot uniaxial pressing (HUP). Mineral waste forms made from clays can be made by FBSR. The clay minerals act as a template: kaolin templates the feldspathoid minerals (sodalite and nephe — line), while illite clays template the micas (see Table 6.8).

In terms of modeling the durability of multiphase ceramics, the distribu­tion of the radionuclides amongst the crystalline phases and in any inter­granular glassy phase is important. Referring back to Eq. [6.1], which gives the needed durability vectors for each phase, we see that the second term drops out since the glassy phase should not have glass-in-glass phase separa­tion leaving terms 1, 3, and 4 (Eq. [6.3]) where the first term should have a minimal durability impact unless large concentrations of the intergranular glass exist or large amounts of radionuclides have been sequestered in the glassy phase compared to the ceramic phase.

X Durability= durability (homogeneous) + durability (CAtAAion)

1st term 3rd term

+ durability (accelerated [6.3]

4th term

If there are multiple types of phases present in the ceramic and each hosts a different radionuclide, then there will be durability vectors for the each phase that hosts that radionuclide as shown in the Table 6.6 example for the 129I in glass-bonded sodalite waste forms.


A metal waste form (MWF) has been under development for stabilization of the metallic fuel hulls from spent nuclear fuel processed pyrochemically.

As the spent fuel is chopped, the fuel materials are removed by the pyro — chemical processing and a stainless steel shell (called a hull or cladding hull) is left in the basket of the bath system. The process removes uranium, acti­nides, and most fission products, leaving behind the hulls, fuel alloy material (generally zirconium), and any noble metal fission products (like techne­tium) in the basket [94]. The noble metal fission products remain somewhat adhered to the surface of the stainless steel hulls and the hulls are coated with salt from the salt bath.

The basket is processed to remove the salt and solidify the hulls, alloy, and other metals into a consolidated waste form. The hulls are solidified by melting the metal into a uniform, homogeneous wasteform (1,560°C). Once homogeneous, the metal alloy should cool to a single phase. Typically, some zirconium (in addition to that remaining from the alloy) is added to bring the metal to about 15 wt% zirconium and lower the melting point of the mass. With the exception of the zirconium to control melting temperature, very few additives are made to the primary waste (cladding hulls), and the overall waste loading is typically above 90% [95].

The metallic waste seems to be a simple waste form with little develop­ment necessary. It has high waste loading, is durable, and fairly straightfor­ward to process. The only development that might make a difference would be an evaluation of whether the cladding could be removed from the process before electrorefining and disposed of separately as a low-level waste form that is potentially greater than Class C. However, the cladding is the host form for the noble metal fission products (notably technetium), and separate disposition would probably require developing a different waste form for those radionuclides.

Likewise, MWF are under study by ANSTO for applications in the United Kingdom by HIPing. In this case, metal encapsulation is to be used for immobilizing debris waste streams that are uneconomical to handle separately, e. g. cermets, SiC, graphite, broken fuel pins, fuel hulls, etc. The process is the same as that used to make glass-ceramic and full-ceramic waste forms and so the processing method is multipurpose.

In spent nuclear fuel (SNF), epsilon metal (a-metal) composed of Mo — Tc-Ru-Pd-Rh is generated from the fission process and heat. The a-metal phase in SNF forms in the same manner that a-metal formed in the natural reactors in Gabon, Africa some 2 billion years ago and has survived largely unchanged except for the decay of 99Tc. Therefore, a-metal has shown long­term stability in nature. This metal does not dissolve during the acid dissolu­tion of SNF but forms solid particles with dimensions of -10 pm in the dissolver sludge. This sludge was formed into a monolithic waste form, by arc melting at 1,800°C into an alloy pellet containing Ru, Re (substitute for 99Tc), Mo, Pd, and Rh in the appropriate masses of each metal [96]. Dissolu­tion rates of 4 x 10-5g/(m2d) and 4 x 10-3g/(m2d) were reported for synthetic £-metal phase and £-phase harvested from SNF under reducing and oxidiz­ing conditions in static durability testing [97, 98].

Добавить комментарий

Ваш e-mail не будет опубликован. Обязательные поля помечены *