Radiation Tolerance of Complex Oxides

See allHide authors and affiliations

Science  04 Aug 2000:
Vol. 289, Issue 5480, pp. 748-751
DOI: 10.1126/science.289.5480.748


The radiation performance of a variety of complex oxides is predicted on the basis of a material's propensity to accommodate lattice point defects. The calculations indicate that a particular class of oxides possessing the fluorite crystal structure should accept radiation-induced defects into their lattices far more readily than a structurally similar class of oxides based on the pyrochlore crystal structure. Preliminary radiation damage experiments substantiate the prediction that fluorites are inherently more radiation resistant than pyrochlores. These results may permit the chemical durability and radiation tolerance of potential hosts for actinides and radioactive wastes to be tailored.

One of the principal factors complicating the selection of materials for nuclear waste storage is the susceptibility of waste forms to detrimental radiation damage effects. Several crystalline ceramics, such as zircon (ZrSiO4) and the orthophosphate monazite (LnPO4; Ln = La, Ce, Nd, Gd, and others in the lanthanide series), exhibit both high chemical durability and solubility for actinides and other radionuclides, and are therefore attractive candidates for nuclear-waste host materials (1,2). However, among the chemically stable host phases proposed for waste storage, there is a paucity of materials for which long-term stability can be anticipated. This is because radioactive constituents in high-level waste (HLW) can decay to produce numerous atomic defects. Most materials are destabilized by such defects, and if defect accumulation is allowed to proceed unchecked, crystalline oxides ultimately succumb to an amorphization transformation, often accompanied by significant volume changes (3) and concomitant microcracking (4).

The technical challenge for HLW storage has therefore been to identify materials for which deleterious radiation effects are averted even at very high self-radiation exposures. The principal consequence of a displacive radiation environment is an elevated population in the lattice of Frenkel pairs (each pair consisting of an atomic interstitial and a lattice vacancy). Subsequent damage evolution hinges on two important factors. First is the degree to which lattice stability is affected by the accumulation of point defects. This factor influences a materials propensity to amorphize under irradiation. The second factor concerns the ultimate fate of irradiation-induced point defects. Interstitials and vacancies can migrate and annihilate harmlessly by interstitial-vacancy (i-v) recombination (the reverse of a Frenkel reaction), or they can cluster with other interstitials and vacancies to precipitate interstitial dislocation loops and voids. A material in which clustering occurs with ease will likely be susceptible to void swelling.

Many simple oxides such as magnesia (MgO) and alumina (Al2O3) are susceptible to void swelling (5). On the other hand, and perhaps surprisingly, a compound made from an equimolar mixture of MgO and Al2O3(best known as the mineral spinel, MgAl2O4) is highly resistant to void swelling under neutron irradiation (5). The radiation-resistant behavior of spinel, which is exceptional for a ceramic, is most likely due to the following factors: (i) Complex chemistry causes the critical size of a dislocation loop nucleus to become unusually large (6). This necessarily suppresses loop nucleation. (ii) Complex structure generates constraints that prohibit dislocation loops from easily unfaulting (7). Faulted interstitial loops remain poor sinks (compared to unfaulted loops) for interstitial absorption. (iii) Some materials like spinel readily accommodate disordering defects within their structures. In fact, the cation sublattices in spinel can be completely disordered by high-fluence neutron irradiation (8). For all of these reasons, harmless i-vrecombination (including cation antisite formation, i.e., swapping the position of one Mg cation with one Al cation) in spinel is a highly efficient point defect annihilation mechanism, and void swelling is negligible.

To test the generality of these attributes for radiation tolerance, we recently initiated an investigation into the radiation damage behavior of an extensive class of complex oxides known as pyrochlores. Oxide pyrochlores are typically ternary compounds of the general formula A2B2O7 (where A and B are metallic cations). The simplest pyrochlore oxides occur in two varieties: (3+, 4+) pyrochlores with the formula A2 3+B2 4+O7; and (2+, 5+) pyrochlores with the formula A2 2+B2 5+O7. For brevity, we concern ourselves here only with (3+, 4+) compounds. Pyrochlores gained notoriety as potential nuclear waste forms because of their structural compatibility with large radionuclide species such as elements in the actinide series (Th, U, Pu) (9).

We used atomistic computer simulation methods to calculate the detailed crystal structure of (3+, 4+) cubic pyrochlore compounds (10). We considered cations ranging from La to Lu on the A site and Ti to Ce on the B site. Our approach allowed us to study not only the perfect crystal lattice, but also to predict the extent to which a given lattice accommodates point defects. Our goal was to develop a quantitative understanding of the trends involved in cation disorder, cation and anion Frenkel disorder, and the interdependence of these disorder mechanisms, as a function of A and B cation radii. Atomistic computer simulation techniques, based on energy minimization with a Born-like description of the lattice (11), were used to generate the various compound structures. Thus, lattice forces are described by effective potentials which have both long-range Coulombic and short-range parameterized components. In addition, the polarizability of ions is accounted for by using a shell model. The model parameters are reported in a study of defects in ZrO2(12). Calculations were carried out using the CASCADE code (13).

We determined the isolated cation antisite defect energy, as a function of A and B cation radii (14), for a wide range of A2B2O7 compounds (Fig. 1). In pyrochlores, the cation antisite is the lowest energy intrinsic disorder mechanism. This mechanism involves the substitution of a 3+ cation onto a B site, i.e., AB4+ 3+(henceforth denoted by A′B), and a 4+ cation onto an A site, i.e., BA3+ 4+ (henceforth denoted by BA ·).

Figure 1

Contour map showing the calculated isolated antisite defect formation energy for a variety of compounds with A2B2O7 stoichiometry. Cations are arranged in order of increasing radii along both the ordinate and the abscissa. For each compound, calculations were performed assuming a cubic unit cell containing eight formula units. Experimentally confirmed pyrochlores are indicated with open symbols; checkered symbols represent compounds known to form fluorites. For compounds labeled with gray symbols, structural data were not available or the compounds have not been observed experimentally. The white region in the lower right-hand corner includes compositions for which the enthalpy of the cation antisite reaction is negative (i.e., cation antisites form spontaneously). Each color contour corresponds to a range of defect energy given in the scale to the right.

Figure 1 provides insight into the stability range of the pyrochlore structure with respect to cation disorder. The plot indicates that antisite defect formation is accompanied by a high energy cost in compounds containing large A cations and comparatively small B cations. The lowest defect energies are associated with compounds in which A and B radii are similar (15). Cation antisite defects are an inevitable consequence of a displacive radiation environment, and Fig. 1 can thus also be interpreted as a predictor of radiation damage behavior: compounds with very dissimilar cationic radii should exhibit the greatest susceptibility to lattice destabilization (and possible amorphization), whereas compounds with similar radii should behave more robustly in a radiation environment (16).

We have included symbols in Fig. 1 to indicate whether a particular A2B2O7 compound is observed experimentally as a pyrochlore. When the pyrochlore structure is not observed experimentally, the fluorite (CaF2) crystal structure is inevitably found in its place (10,17–19). There is a close crystallographic similarity between the pyrochlore and fluorite structures (Fig. 2). The cation sublattices in both cases consist of atoms located at face-centered lattice positions. The only difference is the ordered arrangement of A and B cations in the pyrochlore, compared to the lack of distinction between cations in the fluorite. Similarly, the anion arrangements are identical in both structures, with the exception of a vacant site at an 8a Wycoff position in the pyrochlore lattice (20). Thus, it is not only the ordered arrangement of cations but also the ordered arrangement of anion vacancies which induces a doubly-periodic unit cell in a pyrochlore, compared to a fluorite unit cell. If pyrochlore A and B cations were to randomly exchange places with one another, and anions were likewise to randomly exchange places with anion vacant sites (at 8a), such a pyrochlore would assume the identical periodicity and structure as a fluorite (21). This is precisely what we predict to occur under irradiation. The cation sublattice will disorder via cation antisite reactions, AA +BB → A′B + BA ·, whereas anions will partake in Frenkel defect formation reactions, OO → VO ·· + Oi ", leading to disorder on both cation and anion sublattices (22). The superlattice characteristics of the pyrochlore will be destroyed by irradiation, as the compound adopts the appearance of a disordered fluorite. This effect was recently confirmed experimentally (23).

Figure 2

Schematic drawings comparing the arrangements of cations (A) and anions (B) in the unit cells of pyrochlore (A2B2O7; A,B = cations) and fluorite (MO2; M = cation) compounds (lattice parameters = a). Only one octant of the pyrochlore unit cell is shown in (A) and (B). The various lattice sites in each crystal structure are indicated using Wycoff notation. Pyrochlores and fluorites differ with regard to the ordered arrangement of cations on the pyrochlore cation sublattice and to the ordered arrangement of vacancies on the pyrochlore anion sublattice. Anion vacancies occur in a pyrochlore by virtue of the oxygen deficiency inherent in an A2B2O7 compound, compared to an MO2 compound. A disordered fluorite of composition A2B2O7 consists of a random assemblage of A and B cations of the fluorite cation sublattice (M), along with a random arrangement of oxygen anions (7 per octant) and oxygen vacancies (1 per octant) on the fluorite anion sublattice (O).

All of this discussion is implicit in an analysis of Fig. 1, which indicates that compounds with more similar cation radii are more likely to form as disordered fluorites than as ordered pyrochlores, because the energy expended to form the kinds of defects that cause an ordered pyrochlore to resemble a disordered fluorite (cation antisites and anion Frenkels) is far lower for compounds of similar cation radii, compared with compounds containing A and B cations with highly disparate sizes. Take, for instance, Er2Ti2O7, a compound in the left-hand portion of Fig. 1, for whichr(ErVIII 3+):r(TiVI 4+) ≈ 1.66 (14). Experiments indicate that this compound forms as a highly ordered pyrochlore (24). We also anticipate this compound to behave poorly under irradiation, due to the high formation energies associated with disordering point defects. Apparently, disordering defects destabilize the lattice; this is why Er2Ti2O7 orders upon synthesis. Now consider Er2Zr2O7, a compound located farther to the right-hand side in Fig. 1, for whichr(ErVIII 3+):r(ZrVI 4+) ≈ 1.39 (14). We synthesized a single crystal of Er2Zr2O7 for this study and found that it crystallizes not as an ordered pyrochlore, but as a disordered fluorite (24). On the basis of its structure, we would not expect Er2Zr2O7 to be adversely affected by irradiation; the as-synthesized structure establishes that Er2Zr2O7 is stable in the presence of disordering defects. The structure of Er2Zr2O7 should therefore persist in a radiation environment.

As a test of the predictive capabilities of our atomistic simulation results in terms of radiation damage behavior, we performed ion irradiation experiments to determine the radiation performance of Er2Ti2O7 and Er2Zr2O7. The results verify our predictions (Fig. 3). The Er2Ti2O7 is amorphized by irradiation with heavy ions (Xe) at a fairly low ion dose, whereas Er2Zr2O7 remains crystalline to a high dose of Xe ion irradiation, with no apparent change in crystal structure. These results imply that the zirconate, which commenced existence as a disordered fluorite, is less perturbed by the introduction of defects due to irradiation than is the titanate, which began as a highly-ordered pyrochlore. In some sense, this comes as little surprise because even before exposure to ions, the zirconate resembled an irradiated compound.

Figure 3

Cross-sectional transmission electron microscope (TEM) bright-field (BF) images obtained from (A) a Xe ion–irradiated Er2Ti2O7 single crystal and (B) a Xe ion–irradiated Er2Zr2O7 single crystal. The sample in (A) was irradiated under cryogenic conditions (T ∼ 120 K) using 350-keV Xe++ions to a fluence of 1 × 1015 Xe/cm2. The sample in (B) was irradiated using the same conditions as in (A), but to a higher fluence of 1 × 1016 Xe/cm2. The BF image and the corresponding electron microdiffraction patterns (right) in (A) indicate that the irradiated layer (∼100 nm thick) in Er2Ti2O7 is completely amorphized by the ion irradiation. The BF image and corresponding electron microdiffraction patterns (right) in (B) indicate that the irradiated layer (∼130 nm thick) in Er2Zr2O7 remains crystalline with no change in structure, even at ten times the ion exposure used to irradiate the titanate. In (A) and (B), the incident ion direction (indicated by arrows) coincides with the top of the TEM-BF image.

Many additional experiments support our conclusions. Several radiation damage studies on Gd2Ti2O7 have demonstrated that this titanate pyrochlore is very susceptible to amorphization (23). Also, a recent study on a series of compounds with the composition Gd2(Ti1–x- Zrx)2O2, ranging from Gd2Ti2O7 to Gd2Zr2O7, indicates that radiation resistance improves with increasing Zr content (25). Studies of UO2 (26) and cubic-stabilized ZrO2 (27) have established that compounds with the fluorite structure are especially stable in a displacive radiation damage environment. We propose therefore that complex oxides with pyrochlore and fluorite structures are well-postured to test a new supposition regarding radiation damage effects in ceramics, namely that a material that possesses both complex chemistry and complex structure, and exhibits an inherent propensity to accommodate lattice disorder, should be able to resist lattice instability and possibly void swelling (28) in the presence of a displacive radiation environment. Calculations and preliminary experiments pertaining to a variety of A2B2O7 compounds, with structures ranging from ordered pyrochlores to disordered fluorites, seem to confirm this hypothesis. This information is invaluable to the development of new, chemically durable and radiation-tolerant hosts for safe and reliable storage of radioactive wastes and surplus actinides.

  • * To whom correspondence should be addressed. E-mail: kurt{at}


View Abstract

Stay Connected to Science

Navigate This Article